Abstract
We aimed to analyze the correlation of perineural invasion on transrectal ultrasound guided prostate biopsy with predictors of biochemical cancer recurrence, as well as its impact on clinical outcomes, for non-metastatic prostate cancer. For the study, patients with perineural invasion (N = 86) were recruited into group I and underwent open retropubic prostatectomy, regardless of clinical stage; cases with prostate cancer but without perineural invasion on biopsy, who received radical prostatectomy as the treatment modality, were placed into group II (n = 90). Perineural invasion was detected preoperatively in 43% of cases that revealed surgical margin positivity postoperatively, while 85% of the remaining cases (group II) had negative surgical margins. There was no correlation on prostate biopsy between perineural invasion and Gleason score or PSA, based on Sperman’s rank-order correlation analysis. However, there was strong positive correlation of perineural invasion with clinical stage and patients age. Additionaly, we demonstrated that perineural invasion on biopsy is a non-independent risk factor for metastatic occurrence, although the correlation was significant in univariate analysis. Nevertheless, we found strong correlation between invasion on initial biopsy specimen with biochemical cancer recurrence, suggesting that perineural invasion on prostate biopsy is a significant predictor of worse prognostic outcome.
Introduction
Perineural invasion (PNI) is identified as carcinoma cells along or around a nerve in the perineural space [Citation1]. The clinical significance of PNI remains controversial. It has been suggested that PNI is a predictor of extraprostatic extension (EPE), which is the local spread of prostate cancer (PC) beyond prostate boundaries [Citation2]. PNI could also be a strong predictor of worse oncologic outcomes, including biochemical cancer recurrence (BCR) [Citation3]. However, the independent value of PNI as a predictor of tumor stage has not yet been established. Further, mandatory reporting of PNI differs among countries [Citation4]; the Royal College of Pathologists requires reporting of PNI, but the College of American Pathologist suggests it is optional [Citation4,Citation5]. Although extensively evaluated in radical prostatectomy (RP) specimens, where it has been presumed that PNI relates to worse prognostic factors, the predictive significance of PNIs in prostate biopsy specimens (PBs) is understudied [Citation2].
The aim of our study is to analyze the correlation of perineural invasion on diagnostic prostate biopsy with predictors of biochemical cancer recurrence and age of the patients, as well as its impact on clinical outcomes, for non-metastatic prostate cancer.
Methods
Institution approval for this study was granted by Medical faculty, University of Montenegro. We searched in electronic database of Clinical center of Montenegro for PB pathology reports between 2010 and 2014 that were positive for PC and had cores with evidence of PNI. All PBs were performed under transrectal ultrasound guidance (TRUS) according to the standardized 12 core biopsy protocol [Citation6]. Patients with nonmetastatic clinically localized or locally advanced disease who underwent definitive primary (curative-intent) treatment were included in this study. Conventional staging was performed using computer tomography of the abdomen and pelvis, and bone scan. Treatment occurred within 6 months of the PB date and consisted of RP with or without pelvic lymph node dissection or adjuvant therapy. Pathological specimens were examined, and PNI was identified as carcinoma tracking along or around a nerve in the perineural space [Citation1]. We used the S 100 protein immunohistochemistry assay in the detection of nerves (test No2004127, ARUP laboratories, Salt Lake City, UT). Specimens were examined independently by three experienced pathologists. Nerves were counted in the area of closest proximity of the tumor to the dorso-lateral margins. Infiltration of nerves was categorized on a scale of 0–3; nerves without immediate tumor-cell-contact were given a score of 0, while nerves fully surrounded by tumor (classical perineural carcinosis) were given a score of 3.
Cases were sorted into one of two groups. Group I included all PNI positive patients (n = 86), regardless of clinical stage (CS), who underwent open RRP with median follow-up of 42 months. These individuals were assessed for the presence and intensity of PNI (unifocal and multifocal, category 0–3) in prostate specimens. Cases with PC but without PNI on prostate biopsy, who received RRP as the treatment modality, were placed into group II (n = 90). Since PNI was not detected in this group, categorization of nerve infiltration was not performed. The remaining cases (n = 24) with confirmed PC were treated with radiotherapy (RT) or active surveillance (AS), according to the patients’ preference; Every patient with positive surgical margin (PSM), PNI, or extraprostatic extension (EPE) on definitive specimen after RRP received adjuvant therapy. Pathological specimens of PB and RRP within group I were separately examined for the parameters of PNI, Gleason Score (GS) and CS. For patients undergoing RRP, final pathologic data also included surgical margin status (SMS), presence of lymphovascular invasion, presence of EPE, and delivery of adjuvant treatment. Relevant demographic and biopsy clinical parameters were collected, including patient age, serum PSA (ng/mL) at diagnosis, clinical T stage (TNM classification, 8th edition, 2017), biopsy GS grade group, number of cores involved with PC, and number of cores involved with PNI. [Citation2]. After RP, PSA levels were measured and a DRE was performed at 3-month intervals for the first year, at 6-month intervals for the second year, and annually thereafter [Citation7]
The primary outcome is correlation of PNI on PBs with PSA, GS, CS and patients age. The secondary outcome is the occurrence of metastasis and biochemical recurrence rate (BCRr) among two cohort groups and its correlation with PNI on PBs. Biochemical recurrence is defined as two separate PSA values > 0.2 ng/mL above the nadir [Citation8].
Statistical analysis
Following the customary methods of statistical description, the Student T test was applied to assess statistical significance. The difference of the obtained values is considered significant when p < .05. We apply Sperman’s rank correlation to determine the correlation coefficient between PNI and other variables. Point-Biserial Correlation analysis is used to determine the significance between variables. Finally, we examine the relationship between PNI with BCR and metastatic occurrence using Cox multivariable proportional hazards models. All analyses were conducted using SPSS v.23.0 (SPSS, Chicago, IL, USA).
Results
We identified 200 patients diagnosed with PC, of which 86 (43%) had PNI on PB and fulfilled our inclusion criteria for group I. Ninety patients (45%) without detectable PNI underwent RP, and 24 additional patients were eligible for RT or AS. Demographic and clinical characteristics are shown in . Baseline prognostic variables (PSA, CS and proportion of cores involved with PC) were comparable between the two groups. Clinically, non-organ confined disease was identified in 40% of cases, with a mean of 67% of cores involved with PC. These factors did not differ between two groups (). PSA values ranged between 12 ng/mL (8–20.1) in group I and 10.5 ng/mL (7.3–16.9) in group II. Increase in postoperative GS was observed in 23.2% of the cases in group I. PNI was detected preoperatively in 43% of cases that revealed surgical margin positivity postoperatively; while 85% of the remaining cases (group II) had negative surgical margins. On the final pathology, four patients (4.65%) from group I had stage migration to a higher T stage; none had pathologic understaging (). The median follow-up was comparable between groups (42 vs. 44.7 months; p = .53), during which BCR and metastasis occurred significantly more frequently in the group I (p = .001; p = .03) (). Category 3 PNI was dominant in pathologic specimens, with no significant difference between PBs and RPs ().
Table 1. Demographic and clinical characteristics.
Table 2. Clinical outcomes.
Table 3. Pathologic characteristics of PB and RRP for the first group.
There was no correlation in PB findings between PNI and GS based on Sperman’s rank-order correlation analysis. In order to calculate the correlation between PNI and PSA, we transformed PNI into a dichotomous variable (YES, NO). Point-Biserial Correlation analysis showed that there was no significant correlation between these two variables. Nevertheless, there was strong positive correlation of PNI with CS and age. We found a significant correlation of metastasis occurrence and BCRr with initial PB findings of PNI, but only if the analysis was univariate. (). However, after adjusting for PSA level, clinical stage, and biopsy Gleason score, the relationship between PNI with time to progression was statistically significant only for biochemical recurrence (). Additional adjustment for the percentage of positive biopsy cores, lymphovascular invasion, patients age, or positive surgical margin did not change the results (data not shown).
Table 4. Sperman’s rank correlation for the relationship between biopsy PNI and clinical features (2-tailed).
Table 5. Multivariable proportional hazards models for the relationship between biopsy perineural invasion (PNI) and biochemical cancer recurrence rate (BCRr) or metastatic occurrence, adjusted for standard clinical and pathological variables.
Discussion
Our results reveal a strong correlation between PNI detected on diagnostic PB and the conventional clinical prognostic feature of prostate cancer – BCR, using both univariate and multivariate analysis. Further, we found that metastatic occurrence was associated with PNI in univariate models only. As discussed earlier, the relationship between PNI and biochemical progression after RP [Citation9–11] is controversial. A previous systematic review [Citation12] reported a positive association between PNI with biochemical progression in univariate analysis. Nevertheless, research by Loeb et al. [Citation7] claimed that PNI is a non-independent risk factor for biochemical progression after RP, but is independent of aggressive pathology features (EPE and seminal vesicle invasion) in multivariate analysis. Although their study revealed a significant correlation between PNI and biochemical progression on univariate analysis (similar to our results regarding metastatic occurrence), the relationship was no longer statistically significant after adjusting for PSA level, CS and GS. In terms of the correlation of PNI with adverse pathological and clinical features, our data suggests a strong positive correlation of PNI on PBs with CS and age of the patients; however, no association with GS or PSA was detected. Most patients (85%) had GS 7 or higher on PB, while 20.5% and 9% had GS grade groups 4 and 5, respectively. Analysis of the final RP specimens demonstrated that upgrading and upstaging occurred in 23.2% and 4.65%, respectively, and GS grade 4 occurred significantly more often on final RP specimen (p = .03), in contrast to GS grade 1 (p = .04).
Several studies address the importance of PNI on PBs in pathologic and clinical outcomes of patients with PC. However, results are conflicting, and no definitive conclusion has been made. For example, Elharram et al. [Citation1] claim that PNI on prostate needle biopsy is not predictive for RP outcome and has low predictive value for PNI at RP specimen. However, a majority of studies [Citation13–16] found a strong correlation between PNI and prognostic factors indicating RP and emphasized the importance of PNI in preoperative prediction of the pathologic stage and biochemical recurrence. Katz et al. [Citation13] found that PNI on PBs was strongly associated with a higher pathologic stage; pT3 stage PC was determined in 43.8% of patients with PNI on biopsy compared to 14% of patients without PNI (p = .01). However, all of the prognostic factors identified from the RP specimens were significantly worse in patients with PNI compared to those without. We obtained similar results; we detected a significantly positive correlation between PNI and CS, as well as PNI and patients age, which is comparable to above mentioned study. Loeb et al. [Citation7] found positive association between PNI with EPE and seminal vesical invasion in both univariate and multivariate analyses, but correlation with CS, age of patient, or PSA and GS was not investigated. Furthermore, PNI association with metastatic occurrence was not included in their study. Haki Yuksel et al. [Citation16] claimed that PNI on PBs may be a useful prognostic factor for predicting surgical margin positivity (SMP) in cases with localized prostate cancer. According to their results, SMP was almost 31-fold more frequent in PBs from PNI positive patients, which clearly indicates a significant negative prognostic ability of PNI on biopsy specimens with regards to the advancement of initially localized disease. We also found a significantly higher rate of SMP, lymphovascular invasion, and EPE on final RPs within group I examinees, which confirms a strong predictive value of PNI in biopsy specimens.
The importance of accurate interpretation of pathologic specimens is well established. According to Lubig et al. [Citation17], incomplete perineural involvement (category 1–2) did not have a prognostic value. However, quantitative analysis of the percentage of nerves positive for PNI has a higher prognostic value than identification of PNI alone. In our study group, 57% of biopsy cores were involved with PNI, with majority of category 3 nerve infiltration (44.1 vs. 48.8%). Though we did not estimate the percentage of nerves positive for PNI, it is evident that complete perineural involvement and higher percentage of positive PNI biopsy cores correlate to worse oncological outcome.
Additionally, Goldberg et al. [Citation2] demonstrate that EPE on PB accurately predicts EPE in the RPs with a positive predictive value of 94%. Furthermore, EPE predicts locally advanced disease on final RP pathology and is associated with other PC high-risk features, pointing to an aggressive disease phenotype [Citation18].
In our study, 100% of patients with PNI on PBs showed evidence of EPE on RPs, suggesting that PNI has absolute predictive value on EPE in final prostatectomy specimen. PNI, therefore, has its own prognostic impact on disease aggressiveness and progression potential. Our data agree with findings from a systematic review concerning PNI as a predictor of EPE of PCA [Citation14]; cumulative analysis demonstrated significantly higher incidence of EPE in patients who had PNI at needle biopsy.
There are no contemporary studies of the metastatic predictive value of PNI on PBs. Recent data [Citation2] point out low predictive value of EPE on metastases rate, although the follow-up period was relatively short, and only one third of RP treated patients developed BCR during follow up. Similarly, our results demonstrate that PNI on prostate biopsy is a non-independent risk factor for metastatic occurrence, although the correlation was significant in univariate analysis. Nevertheless, we found strong correlation between PNI on initial biopsy specimen with BCRr, suggesting that PNI on PBs is a significant predictor of worse prognostic outcome and possibly reduced cancer-specific survival.
Our study is not devoid of limitations. First, as a single-center, retrospective report with a moderate sample size and relatively short follow up period, our findings are by no means conclusive; the association of PNI with BCRr and metastatic occurrence is still a matter of debate. Despite this, we report on the largest series of patients with PNI on PB. Notably, we have a higher percentage of PNI on PB compared to larger, more comprehensive studies [Citation7] (43% vs. 15%). We did not addressoutcomes of RT or AS, or of adjuvant therapy follow-up, since this was not the core topic of our study. Second, we did not estimate cancer specific death, nor progression-free survival rate between the two groups; consequently, the definitive predictive value of PNI as prognostic tool is incomplete. Third, objective assessment of PNI (category 1–3) on pathologic specimens is not unbiased, especially at the level of the apex. However, we attempted to mitigate this bias by assigning three pathologist to independently examined specimens.
In summary, this study represents the largest series to describe unique phenomenon of PNI on PB among a cohort of patients with localized or locally advanced PC disease. We have shown that PNI on PB is associated with high risk clinicopathologic prognostic features, indicating an aggressive disease phenotype. PNI on PB also correlates positively with BCR, and represents a non-independent risk factor for metastatic occurrence after RP.
Conclusion
Evidence of PNI on PB is a frequent finding among aging males with PC. PNI accurately predicts the presence of EPE in the final RP specimen. It is also highly associated with BCR after RP, but shows no significant positive correlation to metastatic occurrence after definitive surgical treatment. Our findings underscore the need for pathologists to identify and report PNI on PB.
Disclosure statement
No potential conflict of interest was reported by the authors.
References
- Elharram M, Margel D, Finelli A, et al. Perineural invasion on prostate biopsy does not predict adverse pathological outcome. Can J Urol. 2012;19:6567–6572.
- Goldberg H, Ramiz AH, Glicksman R, et al. Extraprostatic extension in Core biopsies epitomizes high-risk but locally treatable prostate cancer. Eur Urol Oncol. 2019;2:88–96.
- Ball MW, Partin AW, Epstein JL. Extent of extraprostatic extension independently influences biochemical recurrence-free survival: evidence for further pT3 subclassification. Urology. 2015;85:161–164.
- Ahmad AS, Parameshwaran V, Beltran L, et al. Should reporting of peri-neural invasion and extra prostatic extension be mandatory in prostate cancer biopsies? Correlation with outcome in biopsy cases treated conservatively. Oncotarget. 2018;9:20555–20562.
- Srigley JR, Humphrey PA, Amin MB, et al. Protocol for the examination of specimens from patients with carcinoma of the prostate gland. Arch Pathol Lab Med. 2009;133:1568–1576.
- Siddiqui MM, Rais-Bahrami S, Turkbey B, et al. Comparison of MR/ultrasound fusion-guided biopsy with ultrasound – guided biopsy for the diagnosis of prostate cancer. JAMA. 2015;313:390–397.
- Loeb S, Epstein JI, Humphreys EB, et al. Does perineural invasion on prostate biopsy predict adverse prostatectomy outcomes?. BJU Int. 2010;105:1510–1513.
- Roach M, 3rd, Hanks G, Thames H, Jr, et al. Defining biochemical failure following radiotherapy with or without hormonal therapy in men with clinically localized prostate cancer: recommendations of the RTOG-ASTRO Phoenix Consensus Conference. Int J Radiat Oncol Biol Phys.2006;65:965–974.
- O’Malley KJ, Pound CR, Walsh PC, et al. Influence of biopsy perineural invasion on long-term biochemical disease-free survival after radical prostatectomy. Urology. 2002;59:85–90.
- de la Taille A, Rubin MA, Bagiella E, et al. Can perineural invasion on prostate needle biopsy predict prostate specific antigen recurrence after radical prostatectomy?. J Urol. 1999;162:103–106.
- D’Amico AV, Wu Y, Chen MH, et al. Perineural invasion as a predictor of biochemical outcome following radical prostatectomy for select men with clinically localized prostate cancer. J Urol. 2001;165:126–129.
- Harnden P, Shelley MD, Clements H, et al. The prognostic significance of perineural invasion in prostatic cancer biopsies: a systematic review. Cancer. 2007;109:13–24.
- Katz B, Srougi M, Dall’Oglio M, et al. Perineural invasion detection in prostate biopsy is related to recurrence-free survival in patients submitted to radical prostatectomy. Urol Oncol. 2013;31:175–179.
- Cozzi G, Rocco BM, Grasso A, et al. Perineural invasion as a predictor of extraprostatic extension of prostate cancer: a systematic review and meta – analysis. Scand J Urol. 2013;47:443–448.
- De Lancey JO, Wood DP Jr, He C, et al. Evidence of perineural invasion on prostate biopsy specimen and survival after radical prostatectomy. Urology. 2013;81:354–357.
- Haki Yuksel O, Urkmez A, Verit A. Can perineural invasion detected in prostate needle biopsy specimens predict surgical margin positivity in D’Amico low risk patients? Arch Ital Urol Androl. 2016;88:89–92.
- Lubig S, Thiesler T, Müller S, et al. Quantitative perineural invasion is a prognostic marker in prostate cancer. Pathology. 2018;50:298–304.
- Fleshner K, Assel M, Benfante N, et al. Clinical findings and treatment outcomes in patients with extraprostatic extension identified on prostate biopsy. J Urol. 2016;196:703–708.