Prostate Volume, Size does Matter: Growth Dynamics of the Acini and the Stroma using a “Prostatocrit” Model

The gland can be divided into two distinct zones, the peripheral zone which is composed principally of epithelium and the inner gland which is mainly stroma [2,3]. Using this asymmetry of macroscopic and microscopic differences one can develop a new concept, the “prostatocrit” [4] to model the relative growths of epithelial and stromal components. The peripheral zone will have a higher prostatocrit than the transition zone and this ratio is key to the model. This has proven to be more accurate in prostate cancer diagnosis than traditional methods of PSA densities [4].

The rational for this formula lies in the fact that a normal gland is composed of 70% acinal epithelium and 30% stroma. Most of the acini lie in the periphery by definition. The periphery therefore has to be either 80, 90 or 100% acini. We chose to be conservative and elected for 80%. The rest of the acini must lie in the transition zone. The WGv × 0.7 yields the Whole Gland Acinar Volume (WGav). By definition, 1-WGav yields the Whole Gland Stromal Volume (WGsv). The peripheral zone was attributed a percentage of 80%. 0.8 × PZv yields the Peripheral Zone Acinar Volume (PZav). 1-PZav yields the Peripheral Zone Stromal Volume (PZsv). The WGav-PZav yields the Transition Zone Acinar Volume (TZav). 1-TZav yields the Transition Zone Stromal Volume (TZsv).

Densities
The serum PSA is divided into WGav to yield the whole gland acinar density. The peripheral zone acinar density is calculated by multiplying the WGad times the ratio of the PZav/WGav. The transition zone acinar density is WGad-PZad. To ascertain accuracy of imaging, 547 radical prostatectomy specimens with documented gland volumes from histology were documented. Statistics were performed using med calc statistical software.
We used univariate analysis to monitor trends of each zone and subcomponent and related them to IPSS, PSA, growth, densities and ratio with ageing. We then performed multivariate analysis for prediction of PSA.
The larger darker peripheral zone is packed with islands of acini. The smaller lighter transition zone has a few scattered acini ( Figure  1).
We compared the whole gland volume with MRI measurements to estimate if there were any significant difference.
Bland-Altman plot of MRI vs TRUS in estimating volume.
There were 319 TRUS volumes documented pre operatively with 34 MRI volumes documented pre operatively and compared to radical prostatectomy displacement volumes. There was no significant difference in accuracy between the two imaging modalities (Table 1) (Figure 2).

Results
The age range is from 40 to 84 years with the first quartile at 58 and the third at 68.5 years. The PSA ranges from 0.47 to 83 (clearly an outlier given where the third quartile is) with first quartile at 4.6 and third quartile at 9.1. The IPSS ranges from 0 to 35 with first quartile at 11 and the third at 15. The whole gland volumes range from 10cc to 220cc. the transition zone ranges from 2cc to 156cc. the peripheral zone ranges from 3cc to 104cc. Thus we see less range in the peripheral zone compared to the transition zone despite a similar mean volume (Table 2). Using the Prostatocrit model, we can estimate the key indice, the proportion of acini in the peripheral zone (0.8 × 27.7cc = 22.2cc). From this we can deduce that the remaining volume is stroma (27.7cc-22.2cc = 5.5cc). We estimate the whole acina mass as 0.7 × 57.6cc = 40.3cc. We deduce the transition zone acinal volume by subtracting the peripheral acina volume from the total acina volume (40.3cc-22.2cc = 18.1cc).   -4d). There is no significant increase in the IPSS score with ageing (deterioration of 0.025 IPSS units/cc) (P = 0.56). However, when we look at the deterioration of IPSS with overall gland volume we do see a significant effect of 0.025 IPSS units/cc (P = 0.006). This is due to a highly significant transition zone effect of 0.047 ipss units/cc (P = 0.0004) compared to a non-significant association with the peripheral zone of 0.02 ipss/cc (P=0.355). The peripheral zone, which is mostly acini, is not associated with a deterioration in symptom score. The transition zone which has a greater amount of stroma is associated with a deterioration of symptom score with increasing size. The slope is twice as steep for the transition zone compared to either whole gland or peripheral zone.
However, when one look at the relative acinal and stromal growth patterns, we see the transition zone acini growing ×3 that of the peripheral zone acini (0.6cc/year cf 0.21cc/year) with an even more pronounced difference in the stromal components. The transition zone stroma grows at × 6 that of the peripheral zone stroma which is almost negligible (0.29cc/year vs 0.05cc/year). This prostatocrit insight confirms what would be predicted from clinical and pathological experience. There is a greater relative amount of stroma in the transition zone. The peripheral zone has less than half the stroma of the transition zone (5.5cc vs 11.7cc) and the two zones have similar acini volume (22.2cc vs 18.1cc). Here lies the asymmetry (Table 3) The whole gland volumes and PSA are high (right skewed). There is an increase in PSA of 0.23ng/ml/year. When we compare our PSA values with Oesterlings (Table 3) age groups [25], we see almost twice the level for every age group. Our population appears to be twenty years older. This is a population referred because of elevated PSA.

Copyright  Robinson S
PSA secretion is greater in the transition zone acini than the peripheral zone acini ( 0.134 cf 0.081) despite more acini within the peripheral zone. This is well recognised [26,27] and appears to be due to the paracrine growth factor activity of the stoma on the acini. The TZav ( Figure 6d) has a coefficient of 0.286 which explains 53% of PSA secretion (Table 6) (Figures 6a-6f).
Dividing the PSA into the acinal bulk we get a "true" acinal density. There is a non significant increase in overall density with age whether one considers the entire gland or the acinal bulk itself (P=0.107).     Neither does the overall transition zone density increase with time (P=0.425), despite an increase in acini, no doubt due to variable decrements in PSA production.
However, when accounting for TZ acinal bulk using the prostatocrit, the PSA density increases significantly (P=0.0001).   The peripheral zone density appears to increase with time due to the increased cell number, bulk, of acini. The acinar density shows it to not be significant (P=0.879) and this would be expected as ageing cells fail to maintain their rate of PSA production (Table 7)    As expected all measurements of relative growth show a consistent and well recognised picture. The rate of peripheral zone, overall, acinal and stromal growth, relative to the whole gland, decreases with time and hence gives a negative ratio (P<0.0001). The rate of transitional, whole gland, acinal and stromal, increases with time (P<0.0001) and hence yields a positive ratio (Table 8) (Figures 8a-8f).

Copyright  Robinson S
Only age and the stromal component of the Transition Zone (TZsv) were significant predictors of PSA production. Neither the acini of the transition zone nor the peripheral zone acini/stroma nor whole gland volume were significant.

Discussion
Previous studies all reference the whole gland volume as the important entity. We use a "prostatocrit model (Figure 1 We discuss what we mean by size. There has been an almost universal use to refer to the whole gland volume. The gland is quite We can measure the contribution of each element due to the fact that the epithelial components have a secretory product, the PSA. This yields an "acinar density" which we can use to model growth [4]. Conversely we highlight the importance of the transition zone stromal volume which does not secrete PSA. In our previous study it was shown that none of the traditional zonal densities (whole gland, transitional or peripheral zone) were included in a model to predict high grade cancer or all grades of cancer [4]. However, the prostatocrit peripheral zone acinal volumes were included and so we extend this concept to benign growth of the gland predicting that the transition zone stromal volume will be the most significant element in PSA secretion. We made the growth measurements using TRUS with no significant difference found when comparing MRI measurements for whole gland volumes (Table 1, Figure 2).
The growth, proliferation rate of the gland is under the influence of both hormones and ageing. The evidence is conflicting. Regarding Ki67, there was no significant correlation between ageing and proliferation rates in stroma and epithelium. They concluded that a large whole gland volume is not always associated with a high proliferation rate [5]. However, most studies all refer to the global volume of the gland with no discrimination between the subzones. Longitudinal growth rates of the whole gland, not zones, have been assessed [6] using ultrasound. They found an average increase of 1.6%/year. A baseline volume of 29cc showed a general trend to increase with time from 0.3cc/year for younger men and 0.6cc/year for older men. The bigger the baseline volume, the bigger the yearly increase.
Regarding size and symptoms, the Olmsted study [7] described treatment and relation to an enlarged whole gland volume of 30cc. Others have classified BPH as an overall volume of 20g [8]. They did not find a strong correlation (0.22) between volume and symptoms. A weak relation (correlation coefficient 0.185) was found when using a cut-off of 50mls for symptoms and flow rate [9]. regarding obstruction, it is dependent on three parameters, whole gland volume, maximal flow rate and mean voided volume [10] and this is used to generate a bladder outflow obstruction number.
The Krimpen study [11] used a similar population of biopsy negative men and found increasing accuracy of PSA for increasing size of gland. They used 30cc as a cut-off for dichotomisation of whole gland volume. This gave as good area under the curve as higher volumes.
Whole gland volume and PSA are significantly correlated (0.54) and increase with age [12]. They found no influence of volume on symptoms. Whole gland volume on its own is not useful to estimate disease severity. Whole gland volume and serum PSA have an age dependent log-linear relation in those without cancer. The relation is stronger with increasing age and there is a greater increase in overall volume per unit of PSA leading to a "dilution" effect in context of the whole gland volume [13].
Nevertheless, 42% of the variance of whole gland volume can be explained by PSA and age [14]. This has implications for pharmacological outcomes when the whole gland volume is a prognostic factor of treatment.
Trials such as CombAT [15] have classified glands as being enlarged if over 30cc and concluded that a combination of drugs, Tamsulosin and Dutasteride, are efficacious in treating symptoms. MTOPS [16] examined the effects of Doxazosin to relieve tone and Finasteride to induce epithelial atrophy. The risk of progression increased with increasing baseline PSA and whole gland volume of 40mls. The decrease in volume of 19% in those receiving Finasteride refers to whole gland volume only. The reduction in acute retention and need for surgery was attributed to reduction in overall gland volume. Using the PLESS data it was shown that, by dividing men into differing overall gland volumes, that volume and PSA were predictive of the natural history of symptoms and flow rates [17]. In addition, Finasteride gave a better improvement in bother score than placebo [18] after PSA stratification. PLESS was also used to analyse, by volume and PSA, the risk of acute urinary retention and the need for surgery. The risk was higher in those with high baseline PSA and/ or whole gland volume.
More recently focus has been on zonal volumes using MRI [19]. They used zonal contouring to measure the whole gland and the central and peripheral zones. They found a positive correlation between whole gland and central gland volumes and patient age. No correlation with the peripheral zone. Similarly there was a positive correlation of whole gland and the central gland with PSA and with IPSS. They point out that ellipsoid assumptions in calculating the prostate volume are inaccurate. T2 weighted MRI readily distinguishes between peripheral and central zones. They also document how the peripheral zone is relatively static with ageing. They state that the central gland is the major determinant on BPH and elevation of PSA.
Similar techniques using zonal volumes and adjustment of PSA for whole gland and central gland have improved the diagnostic accuracy and personalised risk of cancer [20]. The level of abnormal PSA is a longstanding problematic issue and leads to over-diagnosis and overtreatment of cancer [21]. The accuracy of early changes in PSA as a predictor of lethal cancer is poor [22].
TRUS is quick, simple and safe and not only is useful for volume measurement but also for cancer diagnosis [23]. On the other hand, TRUS has been criticised for being inaccurate [24].
The model is supported by the finding that only age and TZsv were significant predictors of PSA secretion.
Not included were the following. PZav,PZv,TZav TZv WGv. The peripheral zone is not significant at all and the transition acini has no role. This is consistent with prostatocrit stroma being the chief mediator of BPH and LUTS [26,27].

Limitations
Many of the negative biopsies will have undetected prostate cancer and finding a cohort without this is problematic.
We have not been able to follow up men longitudinally and have had to rely on a cross sectional study. MRI will almost certainly Gr up SM Copyright  Robinson S be able to better quantify the zones. Although we have seen no significant difference in MRI and TRUS for whole gland volumes, we do not know that it applies to the zones. The estimation of acini is an approximation based on a standard normal prostate gland in a young man. The amount of acini is not the only factor governing PSA secretion and the relation between epithelium and stroma is difficult to quantify and this is reflected in our low correlation coefficients as well as others [7,8,9]. This complex relationship lies beyond simple measurement of zonal volumes and acinal asymmetry.

Strengths
The prostatocrit makes logical sense and is based on sound anatomic principles. TRUS can define the zones easily and cheaply and it is readily available although operator dependent. TRUS is a good approximation to the volume even if MRI subsequently outperforms it. Although estimation of acini percentage in the zones is problematic the asymmetry is not. We would predict stronger relationships for the effects of drugs in trials such as MTOPS, PLESS and COMBAT if this model were adopted. Potentially better stratification of patients i.e. those most likely to respond would be those with large transition zone stromal volumes.
We will also be able to compare the different growth dynamics of cancerous glands and we plan to demonstrate this. This model has been used to successfully predict the risk of prostate cancer and is superior to whole gland density and transition zone density [4].

Conclusions
The prostatocrit model is new and intuitive. It is consistent with current theories on the role of the stroma in BPH and lUTS. The transition zone stroma is key to PSA secretion and this was confirmed using the model. This model is logical and illustrates known attributes and trends in growth of each zone into its acinal and stromal components, and may offer a more intuitive framework to gauge the behaviour of the gland. The decrease in transition zone acinar density and the increase in transition zone stromal/acinal volume are the most likely cause of LUTS.