Histopathological Findings in Extended Prostate Biopsy with PSA ≤ 4 ng / mL

Objective: Cancer detection has been reported in up to 27% of patients when lowering the PSA cutoff to 2.5 ng/mL. Although this practice could increase the number of biopsies performed, it also could lead to more frequent detection of significant prostate cancers at an organ-confined stage and/or a less aggressive state. This study describes the incidence of malignancy and tumor characteristics in extended prostate biopsies with PSA ≤ 4 ng/mL. Materials and Methods: Prostate biopsies from 1081 patients where examined, 275 (25.4%) patients had PSA level ≤ 4 ng/mL. Results: Cancer was diagnosed in 32.0% and 35.7% of patients with PSA ≤ 4 ng/mL and > 4 ng/mL, respectively (p = 0.906). The median Gleason score was 7 independent of PSA > or ≤ 4 ng/mL (p = 0.078). The median number of cores positive for tumor was 4 and 3, respectively, for PSA > 4 ng/mL and PSA ≤ 4 ng/mL (p = 0.627). There was a difference in the total percent of tumors involving all cores, 11% and 7% for PSA > or ≤ 4 ng/mL (p = 0.042). Fifty-six patients underwent radical prostatectomy, 12 had PSA ≤ 4 ng/mL. In both groups, a diagnosis of cancer was accurate with no differences in Gleason score, tumor volume or staging for both groups. Conclusion: When PSA is below 4 ng/mL, cancer is detected in a proportion equal to the proportion diagnosed with a PSA > 4 ng/mL, and tumor characteristics are similar between the two groups. Only clinically significant tumors were diagnosed following radical prostatectomy.


INTRODUCTION
Numerous investigators have demonstrated the detection of an increasing proportion of earlystage prostate cancer (CaP) and improvement in biochemical outcome after treatment in the Prostatespecific antigen (PSA) era (1)(2)(3)(4).It is also believed to be at least partially responsible for the recent decline in prostate cancer mortality rates in the US and in some European countries (5,6).

Clinical Urology
Traditionally, a PSA cutoff of 4.0 ng/mL has been used to recommend prostate biopsy (7).However, one third of men with PSA level between 4 and 10 ng/mL and more than one half with PSA greater than 10 ng/mL are found to have cancer that has extended to the surgical margins or to the extraprostatic tissue (8).
When the PSA cutoff level is lowered to 2.5 ng/mL, the cancer detection rate has been reported to be up to 27%.Although a PSA threshold of less than 4.0 ng/mL may increase the number of biopsies performed, studies have shown that it also leads to more frequent detection of significant CaP at an organ-confined stage and/or a less aggressive state with no excessive increase in the detection of clinically insignificant cancers (9)(10)(11)(12).
Another matter of debate is the contemporary strategy of extended prostate biopsy, which increases the number of needle cores from 8 to 13, which is a practice that could lead to a greater detection of clinically insignificant cancers.Conversely studies have shown that this practice is responsible for an increase of more than 30% in cancer detection not related to clinically insignificant cancer (13).
Histopathological findings and tumor characteristics have not been well characterized when the PSA cut-off is below 4 ng/mL in the extended prostate biopsy era.To our knowledge complete data including tumor volume have not been previously reported.The aim of this study was to compare the histopathological findings of extended prostate biopsy and radical prostatectomy in men with PSA levels lower or higher than 4 ng/mL.

MATERIALS AND METHODS
From January 1 st 2005 to October 31 st 2006, 1587 biopsies were examined in our laboratory.All information was available for 1081 patients.The mean age was 61.7 years, median 61 (range 31-93).The mean PSA was 7.43 ng/mL, median 5.5 ng/mL (range 0.3-146.0ng/mL).The mean size of the prostate was 57.6 cm 3 , median 48 cm 3 ranging from 15 to 275 cm 3 , and the mean number of cores taken in each biopsy section was 15.5, median 14, ranging from 6 to 47.
Of the 1081 patients, 275 (25.4%) had PSA levels ≤ 4 ng/mL.The median age was 59 years (range 31-78), the mean size of the prostate was 40.6 cm 3 (SD 21.5) and the median number of cores taken in each biopsy section was 14, ranging from 6 to 27.The characteristics of the patients according the PSA levels are in Table-1.
The reason for the biopsy in the men with PSA under 4 ng/mL was available for only 71 (25.8%) patients.Abnormalities in the digital rectal examination was the primary cause, described in 33 (46.5%) patients, followed by suspicious or pre-malignant (prostate intraepithelial neoplasia (PIN) and atypical small acinar proliferation (ASAP)) lesion in previous biopsies in 27 (38.0%),persistent elevation of PSA in 5 (7.1%), family history of prostate cancer in 4 (5.6%) and cancer previously diagnosed in transurethral resection in 2 (2.8%).
Transrectal ultrasound guided prostate biopsies were routinely processed and examined by only one pathologist (KRL).Diagnosis was classified as: 1) benign; 2) suspicious but not conclusive for cancer, also known as ASAP; 3) PIN, and 4) cancer.When the diagnosis was adenocarcinoma, the Gleason score was used for histological differentiation and the tumor extension was shown by the number of cores positive for tumor and total percent of tumor in all cores seen.
A subset of 56 patients, from the 376 who were found to have cancer, underwent radical prostatectomy at our institution.The pathologic analyses of the prostatectomy specimens were completely sampled as described previously in detail (14).Organconfined disease was defined as tumor that did not extend through the capsule, invade seminal vesicles, or metastasize to lymph nodes.Gleason score was used for grading.The tumor volume was determined as a percentage of the prostate gland involved by carcinoma, as estimated using the grid as described by Humphrey and Vollmer (15) and extrapolated to cm 3 for analysis.Staging followed the TNM 2002 recommendations (16).
The differences between the pathologic features were compared between patients whose cancers were detected at a PSA level between 0 and 4.0 ng/mL and those whose cancers were detected after the PSA level rose to greater than 4.0 ng/mL.Standard statistics, chi-square or Fisher's exact test, and Mann-Whitney test analysis were used to compare the data.
Considering the diagnosis, except for PIN, that was more frequently diagnosed in men with PSA ≤ 4 ng/mL, there was no statistical difference between the diagnosis of benign, ASAP and adenocarcinoma (p = 0.906) (Table - The cancer characteristics were similar for both groups (Table -3).The median Gleason score was 7 for both (p = 0.078), the median of number of cores positive for tumor was 4 and 3, respectively, for PSA > 4 ng/mL and PSA ≤ 4 ng/mL (p = 0.627).Considering the total percent of tumor involving all cores, patients with PSA > 4 ng/mL had a median of 11% versus 7% for patients with PSA ≤ 4 ng/mL (p = 0.042).
Considering the 71 patients who had information about the reason of the biopsy, we studied the characteristics of those 33 who were clinical staged T2 comparing with the 38 where digital rectal examination was normal.Among patients that had abnormalities in the digital rectal examination, cancer was diagnosed in 13 (39.4%),comparing with only 8 (21.1%) in 38 without abnormalities in the digital rectal examination (p < 0.0001).PSA levels were similar for both groups, 2.54 ng/mL for T2 patients and 2.73 ng/mL for T1, as was the Gleason score, mean 6.7 for the T2 and 6.1 for T1.Tumors were larger for T2 lesions, with mean number of cores positive for tumor 3.9 and mean total percentage 13.0%,versus 2.3 cores and 2.7% for T1 lesions.Fifty-six patients underwent radical prostatectomy and the findings are shown in Table-4.Twelve had PSA ≤ 4 ng/mL.There was no statistical difference between Gleason score and tumor volume for both groups of patients.The median Gleason score was 7 for both groups (p = 0.068), and tumor volume was 10% or 3.1 cm 3 and 11% or 4.05 cm 3 for ≤ 4 ng/mL and PSA > 4 ng/mL, respectively (p = 0.689 for percentage and p = 0.639 for cm 3 ).There were no differences between the groups regarding extra-prostatic extension (p = 0.424), seminal vesicles infiltration (p > 0.999), lymph node metastasis (p > 0.999) and positive surgical margins (p = 0.427).One (8.3%) patient was stage pT3 with PSA ≤ 4 ng/mL and 10 were staged at this level (22.7%) with PSA > 4 ng/mL (p = 0.424).
In the group of patients with PSA ≤ 4 ng/mL there was no insignificant cancer as defined by Epstein et al. (17) as a tumor volume of less than 0.5 cm 3 , Gleason score less than 7, and organ-confined.Additionally, one patient was stage pT3a, showing extra-prostatic extension and positive surgical margin.

COMMENTS
In order to minimize economic impact in the health system and maximize the effectiveness of detecting and treating CaP, various studies have aimed to find the best levels of PSA and its variations, especially PSA density and PSA kinetics (18).CaP screening programs have shown that using 4.0 ng/mL as a cutoff results in only clinically significant tumors being detected, and one third of the men treated for radical prostatectomy disease that had progressed beyond the prostate (8).Lowering PSA levels to 2.5 ng/mL seems to better detect organ-confined tumors, enhancing chances of disease-free and overall survival, particularly in younger men (19,20).In association with lowering PSA levels, the current practice of more intensive biopsy regimens could lead to the detection of non-significant tumors.It was the aim of our study to describe histopathological findings in extended prostate biopsy in patients with PSA levels lower than 4 ng/mL.
We have shown that cancer was diagnosed in 32% of patients, which is the same proportion of patients diagnosed with cancer with a PSA > 4 ng/mL, which is the traditional cutoff for prostate biopsy.We also observed that cancer was diagnosed even in patients with a very low level of PSA, below 1 ng/mL, and there was significantly worse disease as PSA levels rose.Malignancy was observed in 29.4%, 28.0% and 39.1% of patients with PSA levels from 1.1 to 2 ng/mL, 2.1 to 3 ng/mL, and 3.1 to 4 ng/mL, respectively.Our numbers were even higher than those reported by the Prostate Cancer Prevention Trial, which  Gleason score is the most important isolated prognostic factor, and we observed no difference in the Gleason score between the groups with PSA lower or higher than 4 ng/mL, which both had a median Gleason score of 7. Furthermore, tumor volume in prostate biopsy has been addressed as a very important predictor of cancer extension and outcome.Multiple measurements have been used, including number of positive cores, total millimeters of cancer amongst all cores, percentage of each core occupied by cancer and total percent of cancer in the entire specimen.The best method for determining tumor burden is not yet clear, but estimating a percentage is easy and has been demonstrated to be a useful predictor of tumor extension and cancer-free survival rate (24,25).In the current study we showed no difference considering the number of cores compromised by tumor, but tumors were smaller when PSA ≤ 4 ng/mL, with a total percentage of 7% against 11% when PSA > 4 ng/mL.Smaller tumors are also more likely to be organ-confined.This was not confirmed for patients undergoing radical prostatectomy where tumor characteristics, including volume were very similar to those with PSA higher than 4 ng/mL.One explanation for this fact is a bias considering the choice of treatment.Urologists could have preferred surgery for those patients with other associated adverse characteristics leading to similar results, mostly taking into account tumor volume.This data needs to be clarified in further series.
The detection of organ-confined cancer when PSA is lower than 4 ng/mL could cause some apprehension in treating "harmless" or insignificant cancer.Insignificant cancer is defined as tumor with Gleason pattern less than 4 or 5, organ-confined and volume less than 0.5 cm 3 (17).Reports of fewer than 10% of insignificant cancers have been published, and our series is in agreement with the literature since we did not find any clinically insignificant cancer.In addition to the low number of patients who underwent radical prostatectomy with PSA ≤ 4 ng/mL, our data show tumors that can not be considered insignificant, with mean Gleason score 6.6, median 7, ranging from 6 to 8. In addition it is known that presence of tertiary Gleason 4 or 5, and the percent of a higher Gleason pattern impact the prognosis of prostate cancer.The mean percent of Gleason pattern 4 for this group of patients was 32%, which means a 30% reduction in disease free survival in 10 years (26).Considering tumor volume, McNeal (27) had found good prognosis for tumors with volume less than 4 cm 3 .The mean tumor volume of our surgical specimens was 3.9 cm 3 , but 33% were higher than 5 cm 3 , with one 10 cm 3 , which could be considered a huge tumor, with a 33% probability of recurrence in 10 years (26).
One limitation of our study was the lack of data of PSA velocity (PSAV).PSAV measurement has been shown to be very helpful, as clinically significant prostate cancer is more likely to be found in men with a rapidly rising PSA.Studies suggest that for men with a total PSA higher than 4 ng/mL, a PSA velocity of 0.75 ng/mL/year is an indication for biopsy.However, in men whose total PSA level is lower than 4 ng/mL, an ideal cutoff has not yet been determined and should range from 0.1 to 0.5 ng/mL/year (28)(29)(30).It has been demonstrated that for each 0.1 ng/mL per year increase in PSA, the likelihood of death from prostate cancer increases 15%.For men with a consistent increase in PSA of 0.35 ng/mL per year or higher, the relative risk of dying of prostate cancer is 5 times higher in the next 2 to 3 decades than for men with lower PSA increases (31).Nevertheless, this weak point may be overcome by the findings recently published by Yu X et al. (32).These authors have shown a correlation between total PSA and PSAV, describing a PSAV of more than 2 ng/mL per year in only 1% and 14% of patients whose PSA total levels were lower than 2.5 ng/mL or between 2.5 ng/mL and 4 ng/mL, respectively, indicating a less aggressive and more curable disease.
In conclusion, our findings show that in the extended biopsy era cancer will be detected in 32% of patients when biopsy is performed with PSA below 4 ng/mL.Gleason score and number of cores positive for cancer are similar to those with PSA > 4 ng/mL.Although cancer characteristics in radical prostatectomy were comparable for both groups as Gleason score, percentage of Gleason pattern 4, tumor volume and staging, patients that undergo biopsy with PSA lower than 4 ng/mL are younger and have smaller tumors in biopsies as measured by the total percent of tumor, and, consequently have better chances of having less aggressive tumors.Because of the small number of patients submitted to radical prostatectomy with PSA ≤ 4 ng/mL, other studies, with larger series are warranted to confirm our findings.

EDITORIAL COMMENT
This is a well conducted study concluding that when PSA is below 4 ng/mL, cancer is detected in a proportion equal to the proportion diagnosed with a PSA > 4 ng/mL, and tumor characteristics are similar between the two groups.These findings are supported by other studies.Krumholtz et al. (1) evaluated the pathologic characteristics of clinical stage T1c prostate cancers detected in the 2.6 to 4.0 ng/mL PSA range and compared them with cancers concurrently detected in the 4.1 to 10.0 ng/mL.The authors found that men detected at the 2.6 to 4.0 ng/mL PSA range had significantly smaller cancer volumes however, no difference was found in the proportion of tumors that met previously published criteria of "clinically insignificant" (organ confined, less than 0.2 cm 3 tumor volume, and Gleason sum 6 or less) or "clinically unimportant" (organ confined, less than 0.5 cm 3 tumor volume, and Gleason sum 6 or less) tumors.Using the lower PSA cutoff point resulted in the detection of a significantly higher percentage of organ-confined tumors.The authors conclude that the use of a 2.6 ng/mL PSA threshold for screening resulted in the more frequent detection of small, organ-confined tumors without over detecting possibly clinically insignificant ones.Obviously, additional studies in larger populations with longer follow-up are needed to confirm these findings.

EDITORIAL COMMENT
Early in the PSA era patients with a serum prostate-specific antigen (PSA) level > 4.0 ng/mL and a normal digital rectal examination (DRE) were recommended to undergo prostate biopsy because of a 20-30% risk of prostate cancer at a pre-specified sensitivity of 95%.The majority of such patients have clinically important cancers and the rate of indolent disease, defined as specimen Gleason score 2-6, no extra-prostatic extension, and no Gleason pattern 4/5 is generally < 20%.Many have argued that a PSA threshold for biopsy of 4.0 is more frequently associated with under-rather than over-diagnosis as rates of non-organ-confined cancer (25-35%) are 2 to 4 times higher than indolent cancers, whereas cancers detected in the 2.6-4.0PSA range are more likely to be organ-confined without substantial differences in the rate of low-grade or indolent cancer.In a longitudinal screening study, a decreased risk of PSA-defined biochemical recurrence was observed for patients treated by radical prostatectomy after lowering the PSA threshold for biopsy from 4.0 to 2.5 (1).As such, a lowering of the PSA level for biopsy to 2.5 has been advocated to increase the detection of clinically significant cancers at a more curable stage, and this had been adopted in the guidelines of some professional societies (2).
The Prostate Cancer Prevention Trial (PCPT) has challenged the validity of any PSA threshold for biopsy as no specific PSA value had sufficient sensitivity and specificity for the detection of prostate cancer to be clinically useful (3).Based on the results of patients who had an end-of-study biopsy without usual clinical implications, there was a continuum of cancer risk at all PSA values.Among patients with a PSA < 1.0, 1.1-2.0,and 2.1-3.0, the cancer detection rate was 9%, 17%, and 24%, respectively and the corresponding proportion of cancers graded as Gleason 7-10 was 11%, 12% and 19% (4).These results indicate that there is no PSA below which the risk of having cancer is zero.
In the current study, Leite et al. report on the biopsy and pathological characteristics of a cohort of patients biopsied with a PSA < 4. Reasons for biopsy included abnormal DRE, prior biopsies showing atypical small acinar proliferation or prostatic intraepithelial neoplasia, persistently elevated PSA and negative prior biopsy, or a positive family history, so that the population studied is not fully representative of the general population usually subjected to opportunistic screening.Nonetheless the findings are illuminating, demonstrating similar rates of prostate cancer in those with a PSA < 4 vs. > 4 (32 vs. 36%), no difference in tumor grade (median score 7 in both groups), and slightly fewer positive cores in the PSA < 4 group.In the small subset of patients who underwent radical prostatectomy, there was no difference in tumor volume, grade, or pathological stage.Surprisingly, and unlike our own experience with similar patients where the incidence of indolent cancers is higher in men with PSA < 4, the authors found no indolent cancers as defined by Epstein's criteria of organ-confined tumors of volume < 0.5 cc and grade < 7.This likely reflects the indications for biopsy in this population and less widespread and repeated screening than in the United States.
What then is the optimal PSA cutoff for recommending biopsy in 2008?The theoretical answer is that the optimal threshold is one that maximizes detection of biologically significant but curable cancers, reduces prostate-cancer-specific mortality, and minimizes over-diagnosis of indolent disease.The practical answer is one that recognizes that PSA represents a continuum of risk that is also impacted by many other factors, and that the best way decides whom to biopsy includes a consideration of all of the relevant factors.At the Cleveland Clinic, we have stopped reporting a "normal" cutoff for PSA on our lab reports and substituted the following: "Published data from the Prostate Cancer Prevention Trial demonstrated that there is no PSA level below which the risk of having prostate cancer is zero.For an individual patient, the significance of a PSA level should be interpreted in a broad clinical context, including age, race, family history, digital rectal exam, prostate size, results of prior prostate biopsy, and use of 5 alpha reductase inhibitors.Considering the high incidence of asymptomatic cancer in the general population that may not pose an ultimate risk to the patient, the decision to recommend urological evaluation or prostate biopsy should be individualized after considering all of these factors."We have encouraged the use of the PCPT risk calculator (available at www.compass.fhcrc.org/edrnnci/bin/calculator/main.asp) as one tool (validated published nomograms for this purpose also exist) to achieve the goal of defining individual risk prior to recommending biopsy.Using this calculator, a 55 year old Caucasian male with a negative DRE, a PSA of 1.5, and no family history of prostate cancer has a 19% risk of having prostate cancer but only a 2% risk of having high grade (Gleason 7 or greater) disease, information that can give the patient a more precise estimate of the risks and benefits of undergoing biopsy before deciding whether to have it done.For a 55 year old African American man with a normal DRE, a positive family history, and PSA of 2.4, the calculator estimates a risk of any cancer at 31% and of high grade cancer at 8%; here the risk: benefit ratio probably justifies biopsy even though his PSA is generally considered below the current threshold.Adoption of this approach outside of the U.S. requires construction and validation of similar models on local populations; ultimately, proof of the utility of PSA screening at all awaits the reporting of large screening trials (the ERSPC and PLCO) currently nearing completion.

Table 1 -
Characteristics of 1081 patients submitted to prostate biopsy between January 2005 and October 2006 considering the PSA level.

Table 2 -
Diagnosis of prostate biopsy in 1081 patients with PSA ≤ 4 ng/mL and PSA > 4 ng/mL.

Table 3 -
Tumor characteristics in prostate biopsies of 376 patients with PSA ≤ 4 ng/mL and PSA> 4 ng/mL.
we have just found, biopsy should be recommended when PSA is higher than the median for that specific age, since almost one third of men will be diagnosed CaP.

Table 4 -
Patient age and tumor characteristics in radical prostatectomy specimens when PSA was ≤ 4 ng/mL or > 4 ng/mL.