DISCUSSION
Prostate adenocarcinoma is the most common cancer in men and the second most common cause of mortality after lung cancer. According to 2020 cancer statistics data of the American Cancer Society, prostate cancer constitutes 1 out of every 5 newly diagnosed cancers in men. Among all cancers, it has the highest survival rate with 98%. The reason behind the increased incidence of prostate cancer in recent years is thought to be the start of PSA screening (1).
Almost 75% of the patients diagnosed with prostate cancer are 65 years and older, but it can also be seen in younger people (17). The average age in our study was 63.48 (±6.72).
The correlation between the biopsy GS and the RP GS matters because it affects the treatment method and consequently the prognosis. In the study by Khoddami et al., the biopsy GSs and the RP GSs were the same in 68.2% of the cases, while 1 or 2 GSs differed in 31.8%. The cases had a lower grade in RP in 9.1% whereas 22.7% had a higher grade (18). In another similar study, there was a 36.7% discordance between the scores. It was concluded that the causes of this discordance may be prevented by taking bigger biopsies (19). Our study revealed a similar result with 61.4% of the cases having the same biopsy and RP scores with a higher RP score in 31.2% and a lower RP score in 7.4%. According to WHO-ISUP 2016 classification, it has been argued that Gleason scores 2-5 should not be used anymore due to reasons such as incompatibility between pathologists, the percentage of diagnosis being very low, and the difference between biopsy and RP diagnoses (6).
In the light of new studies that put forth different prognostic results, a new additional GG system has been introduced while the Gleason scoring basically continues with a few HP changes (5–7). With the introduction of the GG system, prognostic research has gained momentum. Mathieu et al. tried to reveal the prognostic value of the new system in a large series of 27122 patients. According to this series, the predicted 4-year RFS for GG I-V was 96.1%, 86.7%, 67.0%, 63.1% and 41.0%, respectively. Although there was no clinical difference in the created multivariate prognostic models, it was stated that the new system could help clinicians and patients to predict the severity of the disease and was easy to use (20). In the study by Epstein et al., great differences were found in terms of recurrence rates between GSs 3+4 and 4+3 and GSs 8 and 9. It was concluded that the 5-tier GG system was able to make the best prognostic distinction in both univariate and multivariate models (21). In our study, when the relationship between the presence of BCR and GGs was investigated, the rate of recurrence from GG I to V was 11.8%, 40.7%, 71.4%, 73.3%, and 94.1%, respectively. This increase was statistically significant (p<0.001) which supports the results of Epstein et al.
The recurrence of prostate cancer after RP is screened by the increase in PSA levels and The American Urological Association has recommended accepting the presence of BCR when two consecutive serum PSA values are ≥0.2 ng/ml. It has been stated that an increase in serum PSA values above this level does not clinically indicate metastatic disease, but it shows disease progression, and follow-up and additional examinations are required to determine clinically metastatic disease (22,23).
In a study of 305 patients investigating the effect of patient age on BCR after RP, no significant relationship was found between age and recurrence (24). Besides, Kunz et al. evaluated the tumor characteristics of patients over the age of 70, which showed that patients in this age group had more biologically aggressive tumors than younger patients, but age alone was not an independent marker of survival after RP (25). Also in our study, no significant relationship was found between age and BCR, and RFS (p=0.974).
Besides GS, some of the HP patterns that have been investigated in terms of their prognostic effects include the presence of EPE, PNI, SVI, LVI, SMP, HGPIN, and tertiary pattern. EPE is an important HP parameter that increases the pathological T stage of the tumor to 3a (26). Ball et al. found that EPE has a significant effect on RFS. They also divided EPE into focal and non-focal groups and concluded that focal ones had worse survival. On top of that, they argued that the pT3a category should also be separated in the future (27). In another study conducted with similar criteria, it was stated that EPE is a significant independent factor in showing BCR in multiple analyzes, and the determination of EPE severity may have positive effects on survival (28). In our study, EPE was detected in 81 (28.4%) patients, and BCR was observed in 44. The relationship between EPE and BCR was statistically significant in single and multiple analyzes (p<0.001) and the presence of EPE was also found to be effective in RFS (p<0.001).
Tumors with SVI are evaluated in the pT3b category and SVI is among the most important determinants of cancer-specific survival, together with the pathological stage after RP (29). Although the number of pT3b tumors has been decreasing with early diagnosis and treatment, SVI is still known to be associated with poor prognostic results (30). The fact that BCR was detected in 29 of 33 (11.5%) cases with SVI in our study proves the bad prognostic effect of SVI.
Although SMP is known to be a predictor of recurrence, it is not certain whether it is a significant indicator of increased risk in cancer-specific mortality (31). In a recent meta-analysis, it was stated that SMP increases the recurrence rate and is an independent prognostic factor (32). The importance of specifying surgical margin positive areas when reporting is not yet known, however, it is among the recommendations to give the length of these areas (33). In our study, SMP was detected in 142 (49.8%) cases, and BCR was observed in 74 of them. The relationship between SMP and BCR was significant in both univariate and multivariate analyzes, and it also had a significant effect on RFS (p<0.001).
PNI is seen in most of the RP cases and studies have conflicting results about its importance. Merrilees et al. did not find the presence of PNI associated with survival after RP, so they did not recommend it to be reported (34). On the other hand, Reeves et al. found that patients with PNI had higher GS, more pT3 disease, more SMP and higher tumor volume, and emphasized the need for more studies to understand the biological behavior of PNI (35). Loeb et al. stated that PNI is an independent risk factor for aggressive pathological features and a dependent risk factor for BCR (36). In our study, PNI was seen in 176 (61.7%) cases, and BCR was observed in 75 of these cases. Although it seemed significant for BCR in univariate analyzes (p<0.001), it was not significant in multiple analyzes (p=0.382).
LVI has been determined as an independent determinant affecting patient survival in a few studies, however, it has been concluded that LVI is not sufficient to show BCR due to insufficient homogeneity in more detailed studies (37). Since it is known that the surrounding stromal tissue can mimic vascular invasion, it is predicted that the evaluation of LVI between pathologists may be different in most studies. Huang et al. concluded that it is an independent prognostic factor in predicting BCR (38). In our study, LVI was observed only in 23 (8%) cases, and BCR was detected in 19 of them. Although it was significant for recurrence in univariate analyzes (p<0.001), it was not significant in multivariate analyzes (p=0.160).
HGPIN is seen as a precursor of prostate cancer. In a study investigating the presence of HGPIN around cancer areas and its importance in RPs, it was observed in 58% of the cases, and it was found that the cancers associated with HGPIN had a better GS, lower PSA value, lower tumor volume and better survival (39). In our study, HGPIN was detected in 183 (64.2%) cases, and it was noteworthy that 86 of these cases had a GS of 3+3, and 61 of them had 3+4. These results support the hypothesis that HGPIN is seen with lower GSs.
The presence of a tertiary pattern reflects aggressive pathological features that show advanced pathological stage and reduced RFS, especially in Gleason score 3+4 and 4+3 tumors (40). Ozsoy et al. found that the cases with a tertiary pattern have a worse prognosis compared to the cases with the same GS without a tertiary pattern (41). In our study, high-grade tertiary pattern was observed only in 19 (6.7%) cases, and it was not statistically significant in terms of BCR (p>0.05).
The results of the studies evaluating the prognostic significance of tumor volume are contradictory. Although tumor volume is known to be associated with pathological stage, studies have not yet shown that it is an independent determinant for survival (42). In a study, it was found that increased tumor volume was associated with SVI, SMP, and lymph node involvement, and it was stated that the increase in tumor volume could be considered as a risk factor after RP (43). In our study, a significant relationship was found between tumor volume and BCR (p<0.001). It was thought that the presence of recurrence in patients with low GS and a low expectation of BCR may be due to the high tumor volume, and the absence of recurrence in patients with high GS and expected BCR may be associated with the tumor volume being 5% or less.
The foamy variant prostate adenocarcinoma, which was seen with a GS of 3+4 in 28 cases and a 4+3 GS in 6 cases in our study, is a subtype with a similar prognosis with the non-foamy type, usually associated with a GS of 7 and pattern 4 (44). Due to the difficulty of its diagnosis, attention should be paid especially in tumors with pattern 4. We couldn’t find a significance in terms of BCR (p>0.05).
RP is still the accepted approach and the ideal treatment for disease-free survival in eligible patients. In their study, Lughezzani et al. stated that even in patients with high-grade cancer with a GS of 8 and above, if the cancer is limited to the organ, long-term oncologic results are quite good (45). Active surveillance is an issue that has been suggested for low-risk, localized prostate cancer, and has been discussed for years. With the increase in early-stage prostate cancer cases as a result of the widespread use of PSA screenings, active surveillance is on the agenda for patients with a GS of 6 and PSA values of <10 ng/ml, by digital rectal examination, periodic biopsies, and serial PSA measurements. It has been shown that this approach is reliable in the long term with the right patient selection, with a cancer-specific mortality rate of 3% in the 10th to 15th years (46). BCR was detected in only 17 of 144 GG I cases with a 3+3 GS in our study, and SMP was observed in 14 of these recurrent cases and inadequate surgery was considered as the reason for recurrence. These data in our study support active surveillance in appropriate and properly selected patients, due to the high levels of RFS in 3+3 patients.
In the latest WHO-ISUP classification, it was recommended to indicate the percentage of pattern 4 in 3+4 and 4+3 cases with a GS of 7, as it plays a role in patient management strategies and active surveillance (6). In a study investigating the effect on RFS by dividing the percentage of pattern 4 into 3 groups (21-50%, 51-70%, ≥70%), pattern 4 percentage was found to be an independent predictor of RFS (47). In our study, the median of pattern 4 was found to be 45% in patients with recurrence and 18% in patients without recurrence, and it was found to be associated with RFS (p<0.05). In the latest Genitourinary Pathology Society (GUPS) White Paper, the preferred method of reporting Gleason pattern 4 percentage was suggested to be either ≤5% or ≤10% and 10% increments thereafter for GGs 2 and 3 (48).
In the latest studies that investigate the different morphological subtypes of Gleason pattern 4, it was reported that cribriform type glands have a worse prognostic outcome in terms of BCR compared to other pattern 4 types (14). In their study in 2015, Kweldam et al. showed that cribriform morphology is a very good predictor of postoperative metastasis and disease-related mortality in tumors with a GS of 7, and in another study in 2017, they found that invasive cribriform and/or intraductal carcinoma after RP was an independent parameter in determining RFS (11,49). Similarly, Choy et al. observed that the presence of cribriform architecture in patients with GS 7 cancer was associated with a reduced 5-year RFS (47). Böttcher et al. found that cribriform and intraductal prostate cancer have increased genomic instability and genomic alterations, and therefore show poor prognostic effects (12). In the GUPS white paper, it was suggested to report the presence or absence of cribriform glands in Gleason pattern 4 (48). After the 3+3 GS cases with BCR detected only in 11.8%, pattern 4 was evaluated in our study according to 4 different histological subtypes. Cribriform glands were observed in 64 cases with Gleason pattern 4. The presence of cribriform glands was statistically significant in terms of BCR in both univariate and multivariate analyzes (p<0.001) and in survival analysis, it was found to be significantly associated with RFS (p<0.001). These results are consistent with the literature and show that cribriform glands have a worse prognostic outcome than other pattern 4 subtypes.
The strong point of our study is the extensive and detailed examination of HP parameters and the patient follow-up data. The limitations are the relatively short median follow-up time and the retrospective design of the study resulting in not being able to perform immunohistochemistry in the discrimination of intraductal carcinoma.