The percentage of affected fragments in needle biopsy in the assessment of pathological staging of prostate cancer

First submission on 26/06/13; last submission on 19/07/13; accepted for publication on 23/07/13; published on 20/10/13 1. Master’s degree in Pathology from Universidade Federal de Pernambuco (UFPE); urologist; volunteer doctor at Hospital das Clínicas (HC) of UFPE. 2. Doctorate in Surgery from UFPE; associate professor at the Department of Surgery of UFPE. 3. Graduating student of Medicine at UFPE. 4. Medical Residency at HC-UFPE; substitute professor of the Department of Surgery at UFPE. 5. Doctorate in Surgery from UFPE; urologist at HC-UFPE. 6. Doctorate in Nutrition from UFPE; full professor of Anatomical Pathology at UFPE. ABStrACt

Preoperative parameters have been used in the selection of patients for surgery. Partin et al. (17) reported that preoperative PSA level, biopsy Gleason score and clinical stage are independent predictors of the final pathological stage. Based on these data, they developed a preoperative nomogram for risk of extraprostatic extension, seminal vesicle invasion and lymph node metastasis. Using the same parameters, D'Amico et al. (4) defined three risk groups concerning biochemical recurrence after RP.
More recently, some studies have shown that measurement of the amount of cancer in prostate biopsies has a predictive value for adverse pathology and risk of biochemical failure (2,5,7,8,19) . The objective of this work is to evaluate the association between the percentage of positive prostate biopsy fragments and the pathological staging of the surgical specimen. Patients who had received adjuvant radiotherapy or hormone therapy, those with stage T1a or T1b, and those with incomplete data in the medical records were excluded from the study. Age was measured by the number of completed years of age at surgery date. The highest serum PSA level before the prostate biopsy was considered the preoperative PSA level. Gleason score was used to assess the histological differentiation of the tumor, and the clinical and pathologic staging was determined according to the Classification of Malignant Tumours (TNM) 1997 -American Joint Committee Cancer (AJCC) (18) .

MAtEriALS AnD MEthoDS
The fragments of prostate biopsy and surgical specimens were completely processed and referred to a pathologist specialized in the field of uro-oncology. Extracapsular extension was defined as invasion of the adipose tissue or the periprostatic neurovascular bundle, classified as pT3a according to TNM system; organ-confined disease, as stage pT2. Seminal vesicle invasion was defined as infiltration of tumor cells into the wall of the seminal vesicle, being classified as pT3b. Positive surgical margin was defined as presence of tumor cells at the inked surface of the specimen.
The percentage of affected fragments (PAF) in prostate biopsy was calculated by dividing the number of positive fragments by the total number of biopsy fragments, and multiplying the result by 100.
In the statistical analysis, we used Student's t-test to study averages, chi-squared test to compare proportions among groups, and multiple logistic regression to analyse the predictive value of PAF. Statistical significance was defined as p < 0.05, and the statistical calculations were performed using software BioStat version 5.0 for Windows.
This study follows the principles of Resolution nº 196, from October 10, 1996, of the National Health Council, on research involving human beings. It was submitted to and approved by the ethics committee of the Center for Health Sciences of UFPE.

rESuLtS
One hundred ninety-two patients were submitted to RP between January 2003 and August 2009. Nine patients were excluded for having received neoadjuvant hormone therapy; one for being classified, according to clinical stage, as T1a; and 23 for incomplete data in the medical records. Table 1 shows the clinical and pathological characteristics of the 159 selected cases. Table 2 compared the mean PAF in the groups of patients with intra-and extraprostatic disease, presence or absence of positive surgical margins (M+ and M-), and presence or absence of seminal vesicle invasion (SV+ and SV-). In the three groups, the difference among means was significant.
In Table 3, the PAF was stratified in three groups: lower than 34%, 34%-50%, and higher than 50%. In each group the incidence of extraprostatic extension was studied, as well as seminal vesicle invasion and positive surgical margins. In the group of patients with PAF lower than 34%, we found 7% extraprostatic disease, 3% SV+ and 14% M+. An increase in the identified pathological alterations accompanied the increase in PAF, and this difference was also significant.
By multiple logistic regression analysis, the PAF presented significant predictive value in relation to the pathological findings in the surgical specimen ( Table 4).

DiSCuSSion
In this article, the study population presented similar characteristics to those of other works in literature, in relation to both age group and mean PSA, and pathological characteristics ( Table 1) (1) , being the mean number of fragments per biopsy greater than in other works.
The main factor associated with the biological behavior of prostate tumor is tumor volume in the RP specimen (22) . Several authors have investigated clinical and pathological preoperative parameters of prostate cancer patients for the development of algorithms to predict tumor stage in the RP specimen (17) . Among the reasons for this interest are the good results with other treatment options for these patients, such as external radiotherapy and brachytherapy; the choice for the surgical technique, such as the nerve-sparing or seminal vesicle-sparing surgery (8,10) ; the presence of comorbidities that exclude surgery as a treatment option (21,23) ; and the identification of patients at high risk of tumor recurrence that could benefit from neoadjuvant or adjuvant therapy (14) .
The amount of cancer in the prostate biopsy has been measured by several methods: total length of cancer in the biopsy in cubic millimeters, total percentage of cancer in the biopsy, number of affected fragments, PAF and maximum length or percentage of cancer in a single fragment (1,2,19) .
PAF in the prostate biopsy has been accepted as the most easily reproducible parameter, since it takes little time from the pathologist and may standardize the different biopsy strategies (1,6,25,26) ; in several articles it has been considered an excellent predictor of poor results after RP (4,7,9,16,21) . In a recent study, Quintal et al. (19) evaluated these several methods to predict extraprostatic disease and identified that the maximum length or percentage of cancer in a single fragment was not significant: the total percentage of cancer in the biopsy was the method with the strongest predictive value.
The importance of PAF in the prediction of organ-confined disease was studied by Sebo et al. (21) . When cancer is confined to prostate, the average of positive fragments is 35%, while in individuals with extraprostatic disease the average rises to 55%. In our work, we found similar values, respectively 33.9% and 52.2%. Lotan et al. (14) reported that the average PAF in patients with positive seminal vesicles was 50%, compared to 22.2% in those with negative seminal vesicles; we found, respectively, 60.4% and 36.4%.
This difference was significantly higher in patients with extraprostatic disease and positive seminal vesicles (pT3) (14,21,26) in comparison with patients with organ-confined disease (pT2), revealing a strong association between tumor involvement on biopsy and pathologic staging.
D'Amico et al. (5) observed that the percentage of positive fragments in biopsy is an important parameter to predict intraprostatic disease, showing that, when less that 34% of fragments are affected, 79% of patients have disease confined to the gland, and when the number of affected fragments is higher than 50%, only 43% have the confined disease. In our study we observed that when there are less than 34% of affected fragments, the incidence of organ-confined disease was 93%, and this value is reduced to 57% in individuals with more than 50% of the fragments affected. Dall'Oglio et al. (6) revealed that seminal vesicle involvement was 6.2% in patients with PAF inferior to 25%, and 31.7% in those with PAF higher than 75%. We found 3% for PFA lower than 34%, and 14% for PFA higher than 50%.
Sebo et al. (21) were the first to confirm the predictive value of PAF and affirmed that in patients undergoing sextant biopsies, the presence of more than one affected fragment (that is, PAF > 17%) has a 3-to 5-fold increased risk. Calvete et al. (3) affirmed that a PAF > 75% is considered of poor prognosis for the presence of extraprostatic disease, and that these patients must receive a combined type of treatment, because surgery only would be little effective. Ojea et al. (16) classified patients in low and high risk according to the classification by D'Amico et al. (4) and stratified PAF in these groups, identifying that among the low-risk patients there were those who presented a high PAF, and these had a higher risk of extraprostatic disease. They also concluded that PAF is useful for individuals with clinically localized disease.
Sakai et al. (20) studied 120 patients who underwent sextant prostate biopsy and reported that the best cut-off value of PAF for prediction of extraprostatic disease was 33%. They suggested that its incorporation to existing nomograms could improve their predictive accuracy. In Table 3, we find significant difference between the group of patients with less than 34% of fragments affected and the groups of patients with 34%-50% and above 50%. Specifically in relation to seminal vesicle involvement and positive surgical margins, there is no difference between groups 34%-50% and above 50%. This suggests that the best cut-off point to predict extraprostatic extension is 33%.
Freedland et al. (7) reviewed the records of 1,094 patients submitted to RP from SouthEastern Aerosol Research and Characterization (SEARCH) database and concluded that PAF in prostate biopsy is a significant predictor of adverse pathology and biochemical failure. They also suggested a risk-stratification approach for recurrence, based on PAF: low risk (up to 34%), moderate risk (35%-50%), and high risk (> 50%).
Partin et al. (17) were pioneers in developing nomograms to predict tumor staging, using PSA, prostate biopsy Gleason score and clinical stage. Gancarczyk et al. (8) studied 1,527 patients and described a nomogram that estimated the probability of extraprostatic disease, involvement of seminal vesicles and lymph nodes. In this nomogram, they added PAF and removed clinical staging from the system, claiming that clinical staging has lost importance lately, as more than 60% of patients currently diagnosed with prostate cancer have stage T1c. Furthermore, they studied the sensitivity and specificity of the nomogram in comparison to isolated prognostic factors, and suggested its superiority, but highlighted the necessity of further studies for its validation.
Prostate cancer, over the last decades, has been considered an epidemic; its incidence and mortality are increasing in our country due to the increased life expectancy and improved diagnostic methods, as well as public prevention policies. The study of preoperative parameters to predict tumor pathological staging allows a more effective treatment, and a resulting increase in the chances of cure. In this context, this article aimed at confirming PAF in prostate biopsy as a tool to be considered in the assessment of prostate cancer.

ConCLuSion
Based on the obtained results of data from 159 patients included in this study, and according to other studies reported in the literature, we concluded that PAF in biopsy is an important predictive factor of the pathological stage in the specimen of RP. It must be used in daily practice, along with other already established parameters, to provide better prognostic and therapeutic management to patients.