The relationship between early biochemical failure and perineural invasion in pathological T2 prostate cancer


  • The views expressed in this article are those of the authors and do not reflect the official policy of the Department of Defense or other Departments of the US Government.

Dr J. Endrizzi, Department of Urology, Wilford Hall Medical Center, San Antonio, Texas, USA.


Objectives To evaluate, in patients with pathologically localized prostate cancer, the relationship between early biochemical failure, i.e. an increasing prostate-specific antigen (PSA) level, and perineural invasion (PNI) on final pathology.

Patients and methods The records were reviewed of 171 patients with prostate cancer who underwent prostatectomy at one institution between January 1992 and December 1995. Data on the histology, therapy and PSA level were collected and evaluated.

Results Of the 171 patients with pathologically localized (pT2) prostate cancer, 131 were evaluable; 17 (13%) had a detectable PSA level in the first 5 years after surgery and 63 had PNI in the pathological specimen. Of those with PSA recurrence, 14 had PNI, one had no PNI and in two there was no comment on PNI. In comparison, only 10 of the 17 patients with recurrence had a Gleason sum of geqslant R: gt-or-equal, slanted 7.

Conclusion Perineural invasion seems to be an important predictor of early outcome in patients with organ-confined prostate cancer treated by prostatectomy. In this series it was the most sensitive predictor of biochemical failure. A more detailed pathological evaluation of prostate cancer may allow the clinician to provide closer surveillance and better informed clinical decision-making.


Historically, little credence has been given to the prognostic significance of perineural invasion (PNI) in adenocarcinoma of the prostate. In his original description of his prostate-cancer grading scheme, Gleason [1] categorized several items that apparently had little significance, amongst which was PNI. Byers and Mostofi [2] noted that PNI occurs early in the natural history of prostate cancer but that this feature added more diagnostic than prognostic significance. There has been a recent trend toward multivariate analysis of risk factors in an effort to provide a better prognosis. PNI identified in biopsy specimens has been reported to be a positive predictor of capsular penetration, with 96% specificity [3]. It has also been closely linked to the mechanism of capsular penetration [4]. Thus we reviewed a series of patients with pT2 disease, assessing the relationship between PNI and biochemical recurrence.

Patients and methods

Patients with prostate cancer treated by radical prostatectomy at one institution between January 1992 and December 1995 were reviewed retrospectively; 171 had pT2 disease. Data on the pathology, treatment and PSA values were collected and evaluated. PSA recurrence was defined as any elevation of PSA above an undetectable level. An evaluable documented follow-up was defined as yearly PSA estimates, including at least one in the previous 6 months; 131 patients met these criteria and were included in the final analysis.


Of the 131 patients 17 (13%) had PSA recurrence during a mean follow-up of 49 months; the mean time to recurrence was 34.7 months. PNI was identified in 63 (48%) of the pT2 specimens. Of the 17 patients with PSA recurrence, 14 had identifiable PNI in the pathological specimen; one of the 17 patients had no PNI and in two patients there was no comment about PNI in the pathological record. Ten of the 17 recurrences were in patients with a Gleason sum of geqslant R: gt-or-equal, slanted 7 in the final pathological specimen and nine patients with recurrence had a PSA level of > 10 ng/mL. The sensitivity, specificity and positive predictive values are shown for each of these factors in Table 1.

Table 1.  The sensitivity, specificity and positive predictive value (PPV, all %) for the Gleason sum, PSA level and PNI
Gleason sum geqslant R: gt-or-equal, slanted 7594715
PSA level > 10 ng/mL538537


Of patients with organ-confined prostate cancer who undergo radical prostatectomy, 10–40% will have PSA recurrence. Persistent PSA levels after surgery are invariably associated with recurrent or residual cancer. Biochemical failure in patients with prostate cancer precedes clinical failure by months to years. The significance of the present results is that PNI may be a precursor to capsular penetration and that in some cases, although capsular penetration cannot be confirmed histologically, it has already occurred. Local and distant microscopic disease are then seen during follow-up as biochemical failure. The extent of PNI and factors that influence it then become more important in evaluating and treating the disease.

Epstein et al.[5] recently discussed the importance of extraprostatic extension and surgical margins in the biochemical failure of patients with Gleason sum 7 pathological prostate cancer. Extensive vs focal extraprostatic extension of prostate cancer were evaluated and related to PSA recurrence in patients treated over a 15-year period. The outcome was worse for those patients with extensive capsular extension and surgical margin positivity than for those patients with focal extracapsular extension or organ-confined disease. In that study, the progression-free probability rate was still < 75% at 10 years for patients with organ-confined prostate cancer, rather than the certainty of being progression-free that might be expected with complete surgical eradication. Progression-free probability rates are lower with extraprostatic extension and margin-negative disease, again indicating that the patient still has prostate cancer.

The results of Bonin et al.[6] suggest that PNI is a significant predictor of biochemical failure in their series of 484 consecutive patients with prostate cancer treated by three-dimensional nonconformal radiation therapy. Of patients with a PSA level of leqslant R: less-than-or-eq, slant 20 ng/mL and PNI on biopsy, 39% were free from biochemical failure at 5 years, while of those with a PSA level of leqslant R: less-than-or-eq, slant 20 ng/mL and no perineural invasion, 65% were recurrence-free at 5 years; the difference was statistically highly significant.

Villers et al.[4] evaluated areas of capsular penetration and surgical margin positivity in 176 radical prostatectomy specimens. The histological evaluation showed two discrete areas where prostatic nerve branches left the neurovascular bundle and extended to the prostatic capsule. These were described as a large superior pedicle at the prostatic base and a small inferior pedicle at the prostatic apex. Seventy-eight prostate cancers with capsular penetration were evaluated histologically and in 39 penetration was exclusively perineural. In the remaining 39 cancers with extraprostatic extension, only four showed no areas of perineural space penetration. Additionally, in these 78 cases, there were 39 in which prostate cancer cells were followed along nerves for at least 3 mm; eight followed the nerve for 6 mm and in one cells were traced along the nerve for 9 mm in serial sections. Bastacky et al.[2], in a review of 302 prostate needle biopsy specimens, reported PNI in 20%; they also noted a specificity of 96% in predicting capsular penetration in the subsequent radical prostatectomy specimen.

The science underlying these clinical and histological observations, and the potential significance of PNI as a prognostic indicator, is becoming better defined. Potential paracrine effects of some growth factors, including nerve growth factor (NGF), on the growth and invasive activity of prostate cancer seem logical, and the idea is not new. DeSchryver-Kecskemeti et al.[7] showed that there is NGF in the epithelium of human prostate cancer specimens. Although it was not quantified, areas of prostate adenocarcinoma showed focal and patchy reactivity with antibodies to NGF, whereas perineural spread of the adenocarcinoma showed a fairly uniform reactivity to NGF. Pflug et al.[8] described the presence of receptors for NGF in human prostate epithelium. Geldof et al.[9] described the effects of NGF-B when added to prostate cancer cell cultures; cells from the DUI45 human prostate cancer cell line appeared to become more invasive in vitro in the presence of this factor. Yang et al.[10] immunolabelled prostate cancer cells undergoing apoptosis that were taken from prostate cancer specimens. Histological analysis of the specimens showed more apoptotic bodies in extra-perineural than in perineural areas. These observations support the possibility of paracrine activity that may promote clinically significant growth and invasion of prostate cancer along periprostatic nerves and out of the gland.

The need for better prognostic indicators in prostate cancer has led to the development of multivariate analyses of risk factors, including the Gleason sum, preoperative PSA level, DNA ploidy, clinical stage and tumour volume, among others. The results of the present small series show the potential prognostic significance of PNI. The presence or absence of PNI is commented on routinely during the histological evaluation of biopsy and radical prostatectomy specimens. Its ready availability in the pathology report facilitated the present retrospective analysis. Further quantification of PNI (i.e. extensive, occasional or minimal) is occasionally available. The relatively short follow-up of the present patients may account for the many false-positive patients (i.e. PNI positive but no recurrence). Additional follow-up of all patients may further strengthen the positive predictive value of this histological finding.

The invasive capacity or malignant potential of prostate cancer is becoming less clear. Prostate cancer grades, as designated by the Gleason score, continue to migrate toward a more moderately differentiated histology, with 80% of prostate cancers being moderately differentiated in a recent report by the American College of Surgeons. As these cancers become more alike, the challenge is to differentiate between them. DNA ploidy is currently being used but remains expensive and is not universally applied. The immunohistochemical differentiation of tumours with the same Gleason score in patients with dissimilar clinical courses may be the key to further grade definition. The potential clinical significance of PNI may stimulate further research in this area. Such pathological data, including PNI, may allow the clinician to provide closer surveillance and thus better informed clinical decision-making.


J. Endrizzi, MD, Chief Resident.

T. Seay, MD, Director of Urologic Oncology.