Introduction

Much attention has been paid in the urologic literature to the concept of ‘insignificant’ prostate cancer (CaP). Beginning with Epstein et al.'s1 first description in 1994, there have been over a dozen studies of the subject.2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13 The construct was originally intended to describe cancers that ‘pose no threat to the patient and might be followed up without immediate treatment’.1 Although much data have been analyzed, it has almost exclusively been in radical prostatectomy (RP) specimens. Crucially, the concept of ‘insignificant’ CaP is thus best applied to patients after they have had definitive surgery of the lesion. What obviously is necessary is a metric that allows such classification based on biopsy criteria, before RP. This has proven elusive.

There have been far fewer studies of ‘insignificance’ in terms of biopsy data alone.14, 15, 16, 17, 18, 19, 20, 21 These are the circumstances when decisions are made by physicians and patients with respect to therapy – specifically low risk therapy in many cases. If the construct of ‘insignificance’ on biopsy is not borne out on subsequent histopathologic examination, then low-risk therapy is less appropriate an option.

These data are critically important not only because they may potentially define a cadre of men for whom definitive therapy may be, or should not be, delayed, but also because they define a potential therapeutic bias that could impact on outcomes after therapies designed for low-risk disease. This construct is paradoxically less crucial when patients are treated with active surveillance22 than with interventions such as cryotherapy or radiation therapy as monotherapy. Under active surveillance, intervention is deferred pending more information. However, if patients with ‘insignificant’ biopsy findings but significant disease are treated with therapy commensurate with presumed low risk, results will be inappropriately poor because of this understaging of disease. This report surveys the literature on RP correlates of ‘insignificant’ CaP on biopsy, and describes other data currently available for the decision-making process before treating such patients as low-risk.

Materials and methods

Searches of the medical literature were performed using PUBMED and MEDLINE. Key words used included ‘prostate neoplasms’, ‘biopsy’, ‘radical prostatectomy’, ‘insignificant’ and ‘unimportant’. All the initial search and related articles were reviewed for content and inclusion. We retrieved citations discussing histopathologic correlates of the subsequent RP specimen in patients undergoing surgery for ‘insignificant’ CaP on biopsy2, 3, 4, 11, 15, 16, 17, 18, 19 (Table 1). Definitions of ‘insignificance’ on biopsy varied between citations, from disease involving a single core18 to disease of <0.5 mm total core length of Gleason sum ⩽6.17 Specific note was made of ultimate tumor volume within the gland in these cases as well as extracapsular extension (ECE) and frequency of positive surgical margins on subsequent RP specimens.

Table 1 Correlates of ‘insignificant’ CaP from the literature with subsequent prostatectomy specimens
Full size table

Results

Table 1 outlines data retrieved from the nine references and two abstracts presented at the 2006 meeting of the American Urologic Association. In each case, a different definition of biopsy ‘insignificance’ was proposed; each had face validity. All definitions had in common clinical T1c stage as part of the proposed definition. Six used Gleason score and one definition had an initial PSA cutoff.

Inspection of the table reveals that prediction of disease in subsequent RP specimens by minimal biopsy disease is distinctly unreliable. For instance, maximal tumor extent within the gland was noted to exceed 10 cm3 in three series. In the six series that noted ECE (over 200 patients), the median frequency was 10.5%, with a range of 6–29.2%. In the seven series that noted positive surgical margins (over 250 patients), the median frequency was also 10.5%, with a range of 0–24%. Median frequency of subsequent discovery of Gleason grade 4 diseases is 14%, with one report of almost 61%. Frequency of bilateral disease was about 80%.

Discussion

Two major trends in CaP diagnosis in the PSA era have merged to create this clinical dilemma of low-volume disease on biopsy. First, there has been an increase in diagnosis of moderately differentiated lesions.23 Further, there has been a decrease in cancer volume of RP specimens.24 Combined, these have led to an increase in the proportion of patients with low volume, moderately differentiated disease on biopsy.25 Although there is uniform appreciation that significant involvement of a large number of cores equates high-volume disease, the converse is simply not true. Thus, it is clear that a proper appreciation for ‘insignificant’ CaP on biopsy is necessary. In up to a quarter of cases, ECE or positive surgical margins arise from such banal biopsy findings. From Table 1, 80% of such patients have bilateral disease! Such patients are inappropriate candidates for low-risk disease therapy, and the question of how to optimize therapy for such patients is not easily answered.

Steyerberg et al.26 recently updated previous work developing a nomogram for indolent CaP. These data were based on 247 patients; correlation was made between pre-operative characteristics (PSA, gland volume on ultrasound, stage, biopsy grade, and lengths of cancer and noncancer specimens) and the ultimate presence of <0.5 cm3 disease confined to the gland without Gleason 4 or 5 components. Their median value, given their patient population, corresponded with only a 45% likelihood of indolent disease (95% confidence interval 38–52%).

We can optimize clinical management of these patients by using further data beyond those provided by the biopsy specimen. Further biopsies may be performed,27, 28 but clearly are not necessary for the pathologic diagnosis; we maintain that once the diagnosis of cancer has been made, another, more robust, repeat biopsy to determine tumor volume represents an unnecessary procedure. The routine use of saturation core biopsy technique for initial diagnosis is questionably appropriate when more limited techniques will suffice. Some constructs are straightforward: a positive biopsy of the seminal vesicle yields a T3b lesion.29 Irrespective of other findings apparent on biopsy, this mandates therapy for locally advanced disease.

Gleason 4/5 component

Of the common parameters currently available on the biopsy specimen, grade may be one of the first discriminants. All of the series reporting insignificant CaP considered Gleason 4 as an exclusion criterion. The highest Gleason score obtained, particularly the component of Gleason 4/5 disease, correlates strongly with PSA and clinical failure,4, 30 although this finding is not universal. For instance, Wang et al.,14 in a series of 59 RP specimens performed for single positive core biopsies, noted no relationship between Gleason score and clinical stage with ultimate tumor volume within the gland.

PSA level and kinetics

Few data exist on PSA values in terms of subsequent RP specimen ‘insignificance’; Wang et al.14 described that a PSA cutoff of 10 ng/ml discriminated significantly larger tumors than values <10 ng/ml. However, the related metric of PSA density (PSAD) is by comparison murky. Noguchi et al.9 found PSAD to be unrelated to lesion significance in their trial. Goto et al.5 described that a single core of 2 mm or less and a PSA density of 0.1 ng/ml/g predicted unimportant subsequent disease with a 75% positive predictive value and 53% sensitivity. Although Augustin et al.10 reported a P value of 0.025 for PSAD on logistic regression in their sample of 1254 men, the combination of a PSAD<0.15 ng/ml/g and 5% biopsy cancer extent yielded a positive predictive value for insignificant disease at the time of RP to be only 27.2%.

Instead, many investigators have reported preoperative PSA velocity (PSAV) as an important predictor of outcome following surgery. Early in the PSA era, Carter et al.31 found an association between pretreatment PSAV and pathologic stage following RP in a cohort of 20 men. Similarly, Goluboff et al.32 reported on 56 men with CaP and found an association between PSAV before RP and pathologic stage, specifically margin status and seminal vesicle invasion. Together, these reports demonstrate an independent association between pretreatment PSAV, Gleason score, time to recurrence, and biochemical failure. More recently, D'Amico et al.33 reported an association between PSAV and prostate cancer-specific mortality (PCSM) and overall mortality. In that study, 1095 men enrolled on a PSA screening study were treated by RP. On multivariate analysis, men were 10 times more likely to die of CaP if their PSA increased 2 ng/ml in the year before treatment. Further, the association between PSAV and death was found to be independent of pathologic factors found at surgery.

Whether PSAV is a marker of tumor aggressiveness or tumor volume has been more difficult to discern. A surgical series from Brazil retrospectively studied 500 men treated between 1986 and 1999.34 The investigators found that a faster PSAV was associated with a larger tumor volume following surgery. Of note, even in the group with the smallest tumor volume at RP (<20% gland; n=305/500), the mean PSAV was high at 3.5 ng/ml/year.

Patel et al.35 reported on 202 men, the majority having low-risk features, treated at Stanford University by RP between 1989 and 2001 and found a correlation between PSAV and pathologic findings. In their review, patients with a rapid PSAV were more likely to have tumors larger than 1 cm3. A total of 86% of their patients with PSAV>2 ng/ml/year had more than >1 cm3 of tumor volume, in contrast to 74% with a PSAV<2 ng/ml/year (P=0.003). This study included various biopsy cancer volume indices including percent positive cores (PPC) and total linear length of tumor in addition to pathologic findings and found only biopsy Gleason score and PSAV independently predicted outcome.

As further evidence to the value of PSAV, D'Amico et al.36 reported an association between preoperative PSAV and postoperative PSA doubling time (PSADT), an early surrogate for PCSM after treatment. They reported that men with a preoperative PSAV>2 ng/ml/year were five times more likely to have a post-prostatectomy PSADT<3 months, and men with preoperative PSAV⩽0.5 ng/ml/year were more likely to have a PSADT>12 months (associated with rare PCSM). Thus, preoperative PSAV is independently correlated with both poor pathologic and clinical outcomes and may be a marker of biologic aggressiveness independent of tumor volume or biopsy cancer volume indices commonly used to identify potentially insignificant CaP.

Number of cores obtained at biopsy

Data from academic and community practices demonstrate an increased cancer detection rate when increasing the number of biopsies from the standard sextant template first proposed in 1989 by the Stanford group,37 to a 12 or greater number of cores, with emphasis on sampling of the lateral and anterior prostate.38 This has resulted in the widespread, but not complete, adoption of at least a laterally directed ⩾8 core biopsy template.

Master et al.39 tested the hypothesis that increasing the number of biopsy cores was directly correlated with tumor volume. Using a modern cohort of over 300 patients undergoing prostate biopsy followed by RP from 2000 to 2003, they showed that an increased number of prostate biopsies detected smaller tumor volumes. Mean tumor volume was 3.85 cm3 for patients with six biopsies versus 2.04 cm3 for patients with greater than six biopsies (P=0.0009). The percentage of T3 cancers decreased from 31% of those patients undergoing six core biopsy to 19% in those undergoing >6 core biopsies (P=0.01). On multivariate analysis, controlling for PSA, Gleason sum, year of biopsy and year of RP, number of biopsies was a statistically strong predictor of tumor volume (P=0.006). Interestingly, the percentages of clinically insignificant tumors, defined in their study as <0.5 cm3 and Gleason sum ⩽6 was not statistically different between the two cohorts (14 vs 22%, P=0.11). Logistic regression analysis was used to identify independent predictors of clinically insignificant tumors. In this study, it was more likely that clinically insignificant tumors occurred when 33% or less versus greater than 33% of the cores were positive (22 vs 6%, P=0.004), or when only one core was positive (37 vs 11%, P<0.0001). This must be considered in the context that, in general, the usual number of biopsy cores has increased over time and tumors may be smaller because they are being detected earlier with PSA screening.

An equally important issue besides the simple volumetric assessment of CaP is the importance of correct Gleason grade assignment.22 In patients who undergo standard sextant biopsies followed by RP, there is at least a 33% chance of the Gleason score being upgraded, with profound implications for the radiotherapy patient. A patient with a Gleason score of 6 might be offered monotherapy as a treatment, but if a patient is upgraded to a Gleason score 8, they would likely receive EBRT and between 6 and 36 months of hormonal therapy. Such patients might also be identified as being at high risk for disease recurrence and be offered participation in clinical trials. Upgrading has been studied by a number of investigators. In one particularly well-characterized series with whole-mount analysis of the prostatectomy specimens, extending the biopsy template from 6 to ⩾8 cores decreased the upgrading rate from 38 to 22% (P=0.04).40 Further, data recently presented revealed that only 18% of RP specimens in patients with presenting PSA<10 ng/ml contained unifocal disease.41

Future direction

It currently appears that the optimum use of positive biopsy results lies not in strategies that attempt to correlate the biopsy with disease volume, but rather in strategies that attempt to correlate the biopsy with biological aggressiveness of the lesion. To this extent, although in its infancy in clinical decision-making, the use of molecular markers (such as p53, Bcl-2, p16INK4A, p27Kip1, c-Myc, AR, E-cadherin and VEGF) is increasing in its availability on biopsy specimens. These markers may prove to be more predictive of outcome than volume of disease or histological grade, and may potentially be used to select patients likely to benefit from more or less aggressive therapies.42

Conclusions

The clinical scenario of minimal volume, moderate grade CaP on biopsy is increasingly experienced in clinical practice. Currently used pretreatment variables frequently do not capture ‘true’ pathology; thus, therapy based on a low-risk clinical presentation may be inadequate. Until molecular markers become more widely established and available, a prudent therapeutic decision for such a patient also involves consideration of the high-grade component, the number of cores obtained, the serum PSA and the PSAV.