The present study represents one of the largest sample bases with long-term follow-up reviewed for VEGF expression in human prostate carcinoma patients, providing much needed incremental progress in the field of understanding angiogenic factors and modern biomarkers for prostate cancer prognosis, which to date has only been investigated in a limited number of studies. Although VEGF expression has been well-studied in many other malignancies, much controversy exists in the literature for VEGF expression and its prognostic and predictive value for prostate carcinoma
[18–28, 31–34]. The need for further understanding in this area of prostate cancer research is thus urgently needed.
The natural history of prostate carcinoma is one with a wide spectrum, ranging from being relatively indolent where a patient may have a life expectancy similar to the general population, to being aggressive with rapid development of metastases and ultimately death
[35, 36]. Currently, with conventional strategies, clinicians have only a moderate ability to estimate prognosis in patients with newly diagnosed prostate cancer, and subsequently there is uncertainty regarding the optimal patient management. These conventional prognostic indicators include clinical tumor stage, Gleason score, and pre-treatment serum prostate specific antigen (PSA) levels
. The development of novel prognostic and predictive biomarkers is thus crucial to identify patients who may benefit from further specific therapy. A prognostic biomarker furnishes information regarding the patient’s overall cancer outcome, irrespective of therapy, while a predictive biomarker may predict response to a particular treatment
Although commonly used tests for women with newly diagnosed breast cancer include estrogen and progesterone receptors and HER2 status, which have both prognostic and predictive value, comparable molecular markers are not available for men with a diagnosis of prostate cancer. Through these molecular markers, appropriate targeted therapy may be selected for patients at a higher risk of cancer progression who will benefit from treatment and avoiding the side effects of therapy from those who will not benefit. Thus, the search for a biomarker of similar utility in prostate cancer is of great importance. A recent review of classical and novel biomarkers as prognostic factors for prostate cancer has highlighted the poor quality and heterogeneity of studies to date, with generally no conclusive results thus far
VEGF is a potent angiogenic factor involved in tumor angiogenesis and represents a potential therapeutic target. Published data regarding VEGF expression in prostate carcinoma has been conflicting. Previous studies in the literature have reported markedly varied VEGF expression in prostate carcinoma, ranging generally from 40% to 100%
[18–28]. There are limited, even fewer, studies on the prognostic and predictive value of VEGF expression in prostate cancer, also with controversial results
In the present study, which is among the largest studies of VEGF expression in prostate cancer with a long-term follow-up of 12.2 years, we explored the potential value of VEGF as a prognostic and predictive biomarker among men with locally advanced prostate cancer enrolled on RTOG 8610. In this study, we found no statistically significant correlation between the intensity of VEGF expression (0–1) and (2–3) for overall survival, distant metastasis, local progression, disease-free survival, or biochemical failure. The results remain non-significant when adjusted for other covariates. There is also no statistically significant predictive value of VEGF demonstrated comparing the treatment arms.
These results are consistent with recent works such as that by Concato et al.
 who investigated molecular markers of cell cycle regulation and angiogenesis and death from prostate cancer. The authors found that there is immunohistochemical evidence of association of bcl-2, p53, or high microvessel density in prostate cancer biopsy specimens at diagnosis and an increased long-term risk of death from prostate cancer. However, there was no statistically significant association with VEGF. A study by Strohmeyer et al.
 also found the importance of microvessel density. In particular, they noted that a high microvessel density was a negative prognostic factor for tumor progression and had a predictive value higher than the classical characteristics of clinical stage, grade, and PSA in prostate cancer after radical prostatectomy. On the contrary, previous reports in the literature by Peyromaure et al. found that VEGF expression in prostate cancer tissue was associated with the risk of cancer progression after radical prostatectomy. Furthermore, Shariat and colleagues
 reported that preoperative plasma VEGF was independently associated with metastases to lymph nodes and biochemical progression after radical prostatectomy, and Vergis and colleagues
 found that increased expression of VEGF and HIF-1alpha were noted in patients at high risk of biochemical failure. It is also interesting to note that Mori and colleagues
 have recently suggested prognostic value for VEGF-A and VEGF-C expression levels in that higher VEGF-A expression was associated with improved overall survival and high VEGF-C expression was associated with decreased risk of developing clinical recurrence. They examined gene expression levels using quantitative real-time PCR. It also differed from our study in that it represented a post prostatectomy setting while our study population consisted of patients treated with primary radiation therapy. The findings of increased VEGF levels in the castrate disease population may suggest that tumor cells acquire new alterations that enable them to survive in the castrated state (adaptation), or represent the outgrowth of rare, pre-existing cells capable of surviving hormonal therapy (selection). The patient population of RTOG 8610 represents a locally-advanced disease population with bulky tumors. The role of elevated VEGF levels in the castrate resistant population may reflect an adaption response, as these patients are on long-term (life) androgen deprivation therapy. The paper by Mori and colleagues
 evaluated short-term androgen deprivation therapy in a preoperative vs. post-operative setting. A possible difference in the findings may be a function of duration of androgen deprivation therapy in the respective patient populations.
There are a number of confounding factors and limitations that may explain the conflicting results in the literature of VEGF expression studies
[18–28, 31–34]. Firstly, the majority of these studies consist of retrospective analyses of very small sample sizes with short follow-up. Ideally, these studies should have long-term follow-up, such as our present study, to examine endpoints such as overall survival rather than biochemical failure. Secondly, measurement of VEGF levels is often problematic, with issues involving tissue availability, sample collection, tissue processing, and storage techniques, all of which could alter results
. In fact, in a study of renal cell carcinoma, Jacobsen and colleagues
 noted that increased storage time resulted in decreased VEGF expression in the membranes of tumor cells from paraffin-embedded tissue samples. Christensen and colleagues
 suggest that, for reliable and consistent results, the ideal conditions for sample collection and preparation should be identified in the study design phase. Then each aspect of sample collection, processing, and storage should be clearly specified in the standard operating procedure document of the study. Thirdly, the level of sensitivity of commonly used assays may be too low to detect meaningful changes in VEGF expression, as even small changes in tumor VEGF expression may be clinically significant depending on the level of dependence of the tumor on VEGF signaling
. Furthermore, the lack of significance as found in this study may be due to the tissue sample size and homogeneity of the patient group. Lastly, there is no consensus in the literature of a “gold standard” VEGF detection assay, and there is a lack of a predefined, accepted, clinically meaningful “cut-off” point for VEGF expression assay measurements. The authors acknowledge that there are other quantitative methods of evaluating VEGF such as examining mRNA or gene expression levels; however, at the time this analysis was performed, the IHC method was the RTOG “standard” at that time and continues to be widely used. Examining the VEGF receptor may further provide informative information in future studies.
In conclusion, the results of much research in VEGF expression in human prostate cancer to date are conflicting. Most studies are only exploratory or hypothesis generating, with small patient numbers. There are even fewer studies on the prognostic or predictive value of VEGF in prostate cancer. Trials are often fraught with inherent challenges as described above. Euphoria is now somewhat tempered because often the initially reported promising results are not reproducible.
In our present study, we found no statistically significant prognostic or predictive value of VEGF expression for locally advanced prostate cancer. This study is among the larger studies of VEGF expression in prostate cancer, and we urge the research community to avoid the misrepresentation of the literature with a lack of publication of even well-designed large negative studies, a publication bias against negative trials, as the current literature in this area appears to be predominated by only small exploratory positive trials, with a lack of subsequent confirmation with larger, longer prospectively designed trials. Thus, to date, the usefulness of VEGF as a prognostic and predictive factor in prostate cancer remains to be clarified. In this study, we had the opportunity to evaluate VEGF levels in the well-characterized RTOG 8610 patient population. There have been a number of other biomarker studies done in this patient population already, and this study complements what is known for this patient population
[45–51]. However, with the limitations presented, we acknowledge that this secondary analysis of RTOG 8610 will not serve to make the definitive statement regarding whether VEGF is a useful biomarker or not, but reporting on this well-characterized patient population with long-term follow-up and numerous other biomarker publications arising from this population is in our opinion a significant contribution to the current heterogeneous VEGF literature
There is an urgent need to establish multidisciplinary initiatives for coordinating further research in the area of human prostate cancer biomarkers, and ultimately strive towards improving the treatment of prostate cancer patients through better targeted therapy.