• angiogenesis;
  • prostate cancer;
  • biomedical imaging;
  • US;
  • MRI


  1. Top of page
  2. Abstract

What's known on the subject? and What does the study add?

Today, angiogenesis is known to play a key role in cancer growth and development. Emerging cancer treatments are based on the suppression of angiogenesis, and modern imaging techniques investigate changes in the microvasculature that are caused by angiogenesis. As for other forms of cancers, angiogenesis is well recognised as a fundamental process in the development of prostate cancer.

The novelty of this extensive report on angiogenesis in cancer, with particular attention on prostate cancer and the imaging techniques able to detect it, is the new prospective to the subject. In contrast with the other available reviews, this report goes from ‘theory’ to ‘practice’, establishing a clear link between angiogenesis development and imaged angiogenesis features. Once the key role of angiogenesis in the development of cancer and in particular prostate cancer has been fully described, attention is turned to the current imaging methods with the potential to assess the angiogenesis process and, as a consequence, to detect and localise prostate cancer.

  • • 
    As confirmed by all available statistics, cancer represents a major clinical and societal problem in the developed world. The form of cancer with the highest incidence in men is prostate cancer. For prostate cancer, as well as for most forms of cancer, detection of the disease at an early stage is critical to reduce mortality and morbidity.
  • • 
    Today, it is well known that pathological angiogenesis represents a crucial step in cancer development and progression. Comparable with most forms of cancer, angiogenesis also plays a fundamental role for prostate cancer growth.
  • • 
    As a consequence, angiogenesis is an ideal target not only for novel anti-angiogenic therapies, but also for modern imaging techniques that aim at cancer localisation by detection of angiogenic microvascular changes.
  • • 
    These techniques are mainly based on magnetic resonance, ultrasound, and nuclear imaging.
  • • 
    This paper provides a comprehensive review of the available studies on angiogenesis in prostate cancer and its use by modern and emerging imaging techniques for prostate cancer localisation.


androgen receptor


cyclooxygenase 2


dynamic contrast-enhanced MRI


diffusion-weighted imaging


fluorine-18 fluorodeoxyglucose


fibroblast growth factor


hypoxia-inducible transcription factor


matrix metalloprotease


MR spectroscopic imaging


mammalian target of rapamycin


microvascular density


platelet-derived growth factor (B)


positron emission tomography


prostate-specific membrane antigen


single-photon emission computerised tomography


tissue inhibitors of metalloproteases


thrombospondin 1


(contrast-enhanced) ultrasonography/ultrasound


US contrast agent


vascular endothelial growth factor (receptor)


  1. Top of page
  2. Abstract

Cancer is one of the leading causes of death in the developed world [1]. The hypothesis that ‘tumour growth is angiogenesis dependent’ was first stated by Folkman [2]. Today, much evidence underlines tumour dependence on angiogenesis in order to progress [3,4].

Human prostate cancer has attracted great interest not only because it is the most common malignancy and the third leading cause of cancer-related mortality in developed countries [1], but also because of its complex, often special nature.

Tumours consist of cancer cells and host stromal cells. These stromal cells, embedded within a protein-rich extracellular matrix and interstitial fluid, face a hostile metabolic microenvironment characterised by hypoxia and acidosis. During tumour development, tumour outgrowth is usually restricted to no more than 1–2 mm in diameter [5]. In this avascular phase, the tumour is nourished by diffusion of oxygen and nutrients provided by nearby blood vessels [6]. Avascular tumours can reach a steady state, where tumour cell proliferation and apoptosis are in balance and a net increase in tumour volume does not occur, so remaining ‘dormant’. In order to exceed the ‘size limit’, the tumour needs an increased blood supply. This is mainly provided by angiogenesis (blood vessel formation from pre-existing vessels, e.g. capillaries and venules) and vasculogenesis (de novo formation of vessels through incorporation of circulating endothelial precursor cells). The transition from a pre-vascular to a vascularised tumour phenotype is referred to as the ‘angiogenic switch’[7]. Once new blood vessel formation is initiated, tumorigenesis and tumour progression may follow by enabling cell shedding from the primary tumour. Invasive tumour cells can possibly metastasise by reaching vital organs directly, via the blood circulatory system, or indirectly, via the lymphatics [8].

Primary or recurrent prostate cancer can be curatively treated if timely diagnosed. However, the process of diagnosis, screening and staging of the disease is still a controversial issue hampered by current detection limitations [9,10].

The rapid development of new imaging methods and the fusion of anatomical, functional, and molecular data have the potential to enable timely detection and characterisation of prostate cancer.

This review gives an overview of angiogenesis in malignant neoplasia, then focuses on prostate cancer and finally provides an overview of current imaging methods.


  1. Top of page
  2. Abstract

At the molecular level, tumour angiogenesis depends on the activation of the endothelial cells, pericytes or the PI3K/Akt/mammalian target of rapamycin (mTOR) pathway in cancer cells [11]. PI3K and Akt are the key elements of the pathway that links the stimulation of growth factor receptors to the phosphorylation and activation of mTOR. The serine/threonine kinase, mTOR, regulates essential signal transduction pathways, and is involved in the coupling of growth stimuli to cell cycle progression [11].

The angiogenic switch tips the balance in favour of angiogenesis by either promoting the activities of pro-angiogenic factors or by inhibiting anti-angiogenic factors [12].

Pro-angiogenic factors

The first group of pro-angiogenic factors comprises the vascular endothelial growth factor (VEGF) [13], a homodimeric glycoprotein. As its name implies, VEGF is a specific mitogen for vascular endothelial cells stimulating both endothelial cell proliferation and migration [14]. There are several families (VEGF, VEGF-B, VEGF-C, and VEGF-D) of which VEGF is the most important, with several VEGF isoforms secreted and various VEGF receptors (VEGF-R) expressed. In particular, two receptor tyrosine kinases, flk–1/KDR (also known as VEGF-R2) and flt–1 (also known as VEGF-R1), bind VEGF with high affinity [15]. Tumour cells can induce new blood vessels by producing VEGF, which is expressed by most cancer types.

Other important angiogenic families are: fibroblast growth factor (FGF, in particular basic FGFs, FGF and FGF2) [16], platelet-derived growth factor B (PDGF-B) [17], angiopoietins (among these, Angiopoietin-1 and Angiopoietin-2) [16,17], growth-related oncogenes [18], TGFβ[19], and matrix metalloproteases (MMPs) [20].

mTOR inhibition can block angiogenesis by disrupting several signalling pathways including inhibition of hypoxia-inducible transcription factor (HIF) 1α translation (see later), VEGF/VEGF-R, and/or PDGF/PDGF-receptor cascade [21].

Anti-angiogenic factors

There are two groups of endogenous anti-angiogenic factors:

  • 1
    proteins and protein-fragments of naturally occurring extracellular matrix and basement membrane components, and
  • 2
    growth factors, cytokines and other non-matrix-derived proteins that directly repress endothelial cell proliferation and migration.

Important non-matrix-derived inhibitors of angiogenesis are interferons and angiostatin [22]. In the matrix-related angiogenesis inhibitors, family thrombospondin 1 (TSP1) [23], and endostatin play a key role [24]. Both angiostatin and endostatin specifically inhibit endothelial cell proliferation and angiogenesis, and potentially inhibit tumour growth metastasis formation [22,24].

Tumour environment

Tumour environment is dominated by tumour-induced interactions. This environment is composed of immune cells, tumour cells, stromal cells, mast cells, and extracellular matrix. Features of the pathophysiological microenvironment are the heterogeneous microcirculation, high vascular permeability, extension of the interstitial fluid space, acidosis, and lack of functional lymphatics [25].

The tumour microenvironment exerts an important role in tumour progression by modulating the metabolism and fostering tumour growth, progression, and metastasis to distant sites. Pro- and anti-angiogenic factors are not exclusively produced by tumour cells, but also by stromal cells of the tumour microenvironment [25].

Tumours other than survive and disseminate, can also mimic some of the signalling pathways of the immune system, thus down-regulating immune cells anti-tumour functions, which not only fail to exercise anti-tumour effector functions, but also promote the angiogenic switch by secreting pro-angiogenic factors [26]. As a result, tumour escapes from the host immune system through the activation of one or several molecular mechanisms that lead to inhibition of immune cell functions or to apoptosis of anti-tumour effector cells.


Hypoxia is a deficiency in the availability of oxygen. In tissue, it may be a required step for a physiological event, in particular in embryogenesis, but may also be characteristic of certain pathological conditions, e.g. ischaemia and cancer [5]. Cancerous tissue, in general, is reported to possess extensive regions of hypoxia compared with corresponding normal tissue [27].

Low tissue oxygen and nutrients concentrations lead to cellular stress responses initiating metabolic and micro-environmental adaptive changes: hypoxia directly regulates these changes through the activation of HIFs [28]. Between the members of the human HIF family, the key regulator of hypoxia-induced angiogenesis is the transcription hypoxia inducible factor HIF-1, a heterodimer consisting of two sub-units HIF-1α and HIF-1β[27,28].


Intracellular pH in tumour cells is neutral as long as they are not oxygen and energy deprived. However, extracellular pH is low due to tumour hypoxia [29]. In order to keep normal pH inside, tumour cells have efficient mechanisms for exporting protons into the extracellular space, which represents the acid compartment in tumours. Cellular pH can be activated by a series of growth factors also involved in tumorigenesis: low extracellular pH causes stress-induced alteration of gene expression, including the upregulation of VEGF in tumour cells.

Cancer cells split glucose into lactic acid, which is a major reason for tumour acidification leading to substantial proton accumulation.

Another source of protons results from conversion of CO2 and H2O via carbonic anhydrase. Both oxygen tension (pO2) and pH are important determinants of tumour growth, metabolism and response to various therapies, e.g. radiation therapy and chemotherapy.

Tumour vascularisation

There are at least four ways for a tumour to overcome growth limitations imposed by insufficient blood supply and to keep growing by gaining access to an enhanced supply of oxygen and nutrients:

  • I. 
    The most important strategy consists of inducing the angiogenic switch, which initiates the process of angiogenesis (discussed so far). The tumour secretes pro-angiogenic factors and/or suppresses anti-angiogenic factors resulting in the induction of endothelial cell proliferation and migration, vessel sprouting, and tube formation (Fig. 1).
  • II. 
    Vasculogenesis forms the initial vascular network during embryonic development and contributes to subsequent vessel formation in some tissues [30].
  • III. 
    Co-option of existing vessels allows tumours to grow along pre-existing blood vessels and use these vessels for their blood supply [31].
  • IV. 
    Vasculogenic mimicry forms pseudovascular channels by tumour cells rather than endothelial cells [32].

Figure 1. Most tumours develop starting from an avascular phase (a). The angiogenic switch allows for a transition from the dormant phase to the vascularised one by balancing the ratio between pro- and anti-angiogenic factors. As a result, vessel dilation (b) followed by sprouting and tube formation (c) are inducted. Tumour blood vessels supply vascular nutrients and oxygen to the tumour mass as long as it grows (d). Invasive tumour cells may form metastases by entering the blood vasculature after basement membrane degradation. Alternatively, they can penetrate the lymphatic system and be transported to regional lymph nodes (e).

Download figure to PowerPoint

It is worth emphasising that more than one strategy may be found in the same tumour, and tumours might adopt different strategies according to their stage and grade of malignancy.

The assessment of tumour microvascular density (MVD) has been reported as a prognostic indicator in those tumours that induce significant angiogenesis, e.g. breast [4], prostate [33,34] and ovarian cancer types [35], where higher tumour vascularity correlates with worse patient outcomes. In contrast, carcinomas of lung and urinary bladder do not show significant associations of MVD with poor prognosis, reflecting differences in angiogenic mechanisms between different tumour types [36,37].

Vascular architecture

Tumour vascular organisation may be completely different depending on the tumour type, its location, and whether it is growing, repressing, or relapsing. In general, tumour vessels are immature: they are irregular shaped (dilated, saccular, tortuous) and chaotic in their patterns of interconnection (lack a normal hierarchy of vessels types) [5]. Whereas the normal vasculature is characterised by dichotomous branching, the tumour vasculature has arterovenous shunts, many trifurcations and branches with uneven diameters [29]. Moreover, having abnormal or decreased pericyte coverage, tumour vessels leak fluid and blood into the surrounding tissue.

Blood flow and microcirculation

Whether normal or abnormal, arterio–venous pressure difference and flow resistance govern blood flow in a vascular network. Flow resistance is a function of vascular architecture (referred to as geometric resistance) and blood viscosity (referred to as viscous resistance). Abnormalities in both vascular architecture and blood viscosity increase blood flow resistance in tumours, counterbalanced by the presence of arteriovenous shunts and a lack of vasomotor control that cause a low flow resistance. Microvascular architecture disorganisation results in limited blood supply and inefficient oxygen transport. As a consequence, local ischaemia is induced. Furthermore, tumour blood flow is unevenly distributed, fluctuates with time, and can even reverse its direction leaving ischaemic regions. The presence of ischaemic regions contributes to hypoxia [27], and also opposes the delivery of therapeutic agents [38].

Vascular permeability

Diffusion is considered to be the major form of transvascular transport in tumours. The diffusive permeability of a molecule depends on its size, shape, electric (ionic) charge, and dynamics of the transvascular transport pathway. Several pathways have been described for molecule extravasation across the endothelium in normal tissues [39].

Vascular permeability to solute and water (referred to as hydraulic conductivity) of tumours is generally higher (≈4–10 times) than that of various normal tissues [40]. As a result, tumour vessels lack size selectivity of extravasating molecules, leading to extravasation of plasma and plasma proteins into the surrounding connective tissue.

In tumour, high vessel permeability together with lack of functional lymphatics contributes to interstitial hypertension (>15 mmHg vs ≈0 mmHg in normal tissue) [38].

Despite a general increase in permeability, not all the blood vessels of a tumour are leaky. Not only does the vascular permeability vary from one tumour to the other, but it also varies spatially and temporally within the same tumour as well as during tumour growth, regression, and relapse [38].


  1. Top of page
  2. Abstract

Prostate cancer characterisation

Adenocarcinoma, the prostate cancer malignant neoplasm, predominantly arises from epithelial cells in the peripheral zone (≈70% of cases). Usually less affected are the central zone (≈25%) and transitional zone (≈5%) [41].

Angiogenesis is an important process also in prostate cancer progression, being critical to tumorigenicity and metastasis [42]. Prostate cancer has the ability to produce MMPs, VEGF, TGFβ, and cyclooxygenase 2 (COX-2). Several endogenous inhibitors of angiogenesis have also been described in prostate cancer, namely angiostatin, endostatin, PSA, TSP1, interleukin 8, and interferons.

The microenvironment of prostate cancer is a critical determinant in cancer genesis [43]. Bidirectional cellular interactions between neoplastic prostate cancer cells and stroma cells are mandotory for local tumour progression and metastasis, and influence the tumour microvascular architecture.

At present, prostate cancer grade is evaluated by histological Gleason score, a measure of cell differentiation and widely accepted as a pathological indicator of biological behaviour, correlating with stage and metastatic potential. However, Gleason grading of the prostate biopsy remains a poor predictor of pathological outcome [44]. Taking into account the essential role of angiogenesis in prostate cancer developments will most probably lead to further improvements in prostate cancer diagnosis and staging.

Growth factors


Pre-clinical data with prostate cancer cell lines show that VEGF is a potentially important factor in stimulating cell proliferation as well as angiogenesis and lymphangiogenesis [45]. VEGF regulation is mediated by factors such as hypoxia, cytokines, and androgens; it can also involve activation of oncogenes (Ras-, Raf-, and Src) and inactivation of tumour-suppressor genes (p53 and von Hippel–Lindau) [46]. During embryogenesis, lymphatic vessels develop from blood vessels. Therefore, lymphangiogenesis and angiogenesis are stimulated by the same family of growth factor proteins. In particular, VEGF-C, VEGF-D, and their tyrosine receptor in lymphatic endothelial cells (VEGFR-3) are potent lymphangiogenic factors in prostate cancer [47].

VEGF levels, measured by immunohistochemical analysis, appear to be higher in patients with metastatic prostatic cancer than those with localised prostatic cancer, but this evidence does not appear to be useful in predicting prostate cancer progression [48]. Moreover no direct correlation has been found between VEGF expressions and age, serum PSA level, and Gleason score [49].


In prostate cancer, several FGFs are overexpressed [50]. Immunohistochemical studies have shown FGF2 receptor protein expression in prostate cancer cells and in endothelial cells. Low-grade tumours, as well as androgen-dependent prostate cancer cells, do not synthesise significant amounts of FGF2 and express only a small amount of FGF2 receptors. Prostate cancer cells and androgen-independent, more aggressive cells, on the contrary, synthesise a large amount of FGF2 and express many FGF2 receptors [50].

Studies evaluating the usefulness of urinary and serum VEGF and FGF2 measurements have shown no correlation with prognosis and are not as useful as PSA measurements [51]. However, the role of FGFR2 and changes in alternative splicing of this receptor in prostate cancer progression are still unclear.


TGFβ acts in a paracrine mode in benign prostate epithelium to maintain epithelial homeostasis in the prostate [52]. Prostate cancer cells overexpress TGFβ, which in turn promotes carcinogenesis by increasing extracellular matrix production, inducing angiogenesis, inhibiting host immune function and, therefore, enhancing tumorigenicity of prostate cancer [52]. Furthermore, TGFβ was recently shown to favour osteoblastic bone metastases in experimental systems. There are three TGFβ isoforms, namely, TGFβ1, TGFβ2, and TGFβ3 that exert their biological effects by binding to cell surface receptors, TGFβRI, TGFβRII, and TGFβRIII, respectively.

In prostate cancer it is well established that an increase of TGFβRI is associated with high-grade tumours and higher clinical tumour stage.

Additionally, TGFβRI expression correlates with tumour vascularity, tumour grade, and metastasis [53]. TGFβRIII expression is decreased or lost in most human prostate cancer in comparison with benign prostate tissue. This loss correlates with advancing tumour stage and higher probability of PSA recurrence, suggesting TGFβRIII expression to play an important role in prostate cancer progression. Finally, loss of TGFβRII responsiveness in fibroblasts results in intraepithelial neoplasia in the prostate.

Taken together, these studies suggest that partially blocking TGFβ could potentially block a constant stimulation of many downstream factors in the process of tumour angiogenesis.


MMPs contribute to metastases not only by modulating the remodelling of the extracellular matrix, but also by regulating angiogenesis in cancer, both positively, through their ability to activate pro-angiogenic factors, and negatively, through generation of angiogenesis inhibitors, e.g. angiostatin and endostatin, which are cleaved from large protein precursors [54].

The MMPs' activities are subject to regulation by tissue inhibitors of metalloproteases (TIMPs) [55]. An association between MMPs and/or TIMPs expression and parameters indicative of prostate cancer aggressiveness has also been reported. Recently, Escaff et al.[56] reported that MMPs/TIMPs expressions are in general higher in prostate cancer than in prostate benign tissue, confirming the important role of these factors in the molecular biology of prostate cancer.

In addition to the aforementioned factors, androgens and prostate-specific membrane antigen (PSMA) also play a crucial role in prostate cancer tumorigenesis:

  • • 
    Androgens are essential for the development, growth, and maintenance of prostate cancer. The effects of androgens are exerted via the nuclear androgen receptor (AR), a ligand-dependent (either testosterone or 5α-dihydrotestosterone) transcription activator. The AR is involved in diverse cellular functions including differentiation, angiogenesis, proliferation, and apoptosis [57]. Antiandrogen therapy (androgen ablation) is a common treatment for prostate cancer. Recently, links between antiandrogen therapy and angiogenesis have been discovered. Addition of VEGF inhibitors to hormonal therapy could result in increased oxygen delivery to hypoxic tumours areas and thus further potentiate radiation therapy [57].
  • • 
    PSMA expression in tumour-associated neovasculature is necessary for angiogenesis and endothelial cell invasion, although the role of PSMA in angiogenesis has not yet been fully elucidated [58]. Unlike other neovascular targets, e.g. VEGF, which are expressed in both healthy and pathological prostate, PSMA is consistently expressed in BPH and prostatic adenocarcinoma, but not in normal vasculature [59]. This specificity makes PSMA an ideal target. Novel PSMA-based prostate cancer therapies are currently under investigation [58,59].

Endogenous inhibitors

Endogenous inhibitors of angiogenesis (e.g. angiostatin and endostatin) that are expressed by most tumours, also exhibit a key role in prostate cancer angiogenesis [42]. In addition, several anti-angiogenic factors more specific for prostate cancer have been identified:

  • • 
    PSA is secreted by endothelial cells into seminal fluid, and only minor amounts reach the bloodstream. In prostate cancer, PSA expression is reduced and more aggressive tumours have lower PSA levels than less aggressive ones, but the serum levels are elevated due to increased leakage from the prostate. Serum PSA level is therefore a useful biomarker for prostate cancer [60];
  • • 
    TSP1 belongs to the family of extracellular matrix proteins and can inhibit angiogenesis by inhibiting growth of endothelial cells. Normal prostate cells secrete high levels of TSP1, whereas short-term cultures of BPH and prostate cancer cells secrete low TSP1 levels. TSP1 down-regulation correlates to progression in proliferative disease [61];
  • • 
    interleukin 8, a mitogen for endothelial cells, may also play a role in regulating angiogenesis. Prostate cancer cells express interleukin 8, which appears to increase angiogenesis and metastasis through induction of MMP-9 expression [62];
  • • 
    interferons (including interferon α, β and γ) have shown significant antitumor activity in pre-clinical models and are among the most commonly used cytokines in patients [63]. The effects of interferon α and interferon β on the vasculature have been mainly attributed to inhibition of basic FGF production by tumour cells or down-regulation of interleukin 8 and VEGF gene expression. Furthermore, interferon γ up-regulates the surface densities of many molecules and down-regulates the expression of other surface proteins in prostate cancer.


The same as for most types of cancer, there are two distinct angiogenic events involved in the formation and progression of prostate cancer: the first ‘initiation switch’ and the second ‘progression switch’. During prostate cancer evolution, microvessel density and microvascular structure undergo several changes.

Hypoxia-induced angiogenesis

HIF is upregulated in most prostate tumour tissues, compared with normal and benign prostate tissues [64]. Furthermore, hypoxia can up-regulate the expression of VEGF in prostate cancer [65].

Cvetkovic et al.[65] proposed that prostate tumour cells adapt to decreased pO2 levels by transcription of HIF-1, which causes VEGF production, ultimately leading to enhanced angiogenesis.

Zhong et al.[64] linked the signal transduction pathway from receptor tyrosine kinases to PI3K signalling and Akt activation, which turns on mTOR and VEGF-induced angiogenesis in prostate cancer.

Hypoxia also induces COX-2, an enzyme that stimulates angiogenesis through interactions with VEGF. COX-2 overexpression has been detected in the angiogenic vasculature present within the tumours and pre-existing vasculature adjacent to cancer lesions, suggesting that COX-2 may induce newly formed blood vessels to sustain tumour cell viability and growth [66].

Androgens can activate HIF-1 through an autocrine loop which in turn activates HIF-1α and HIF-1-regulated gene expression. HIF-1 interacts with AR on PSA gene promoter, thereby activating its expression [67].

MVD in the prostate

Several investigators have found MVD to be higher in prostate cancer when compared with benign glands. Increased MVD was found in the primary tumours of patients with metastatic disease compared with patients with localised disease [33]. Higher MVD counts have been correlated with advanced pathological stage [68], increased PSA levels [69], higher tumour grade [70], increased metastatic potential [33,70], and decreased survival of patients with prostate cancer [70,71]. However, no or even negative correlations between these endpoints are reported by other groups [72,73].

Some of the controversies may be due to the specific methodology used to assess MVD, e.g. the use of different antibodies, the types of cases examined and, most importantly, the heterogeneity of the vascular system geometry. These geometrical characteristics are source of errors in visual counting. Imaging techniques that have the potential to visualise lesions with increased MVD have been proposed [34,74].

Microvascular architecture

The same as in other types of cancer, prostate cancer vasculature is composed of two types of vessels: the existing vessels in the surrounding normal tissues into which the tumour has invaded, and the tumour microvessels arising from neovascularization.

The neovessels are highly irregular and tortuous, and have arteriovenous shunts, blind ends, and basement membranes with no smooth-muscle layer. In particular, going from BPH through prostatic intra-epithelial neoplasia up to invasive adenocarcinoma, an increasing proportion of capillaries become smaller, shorter, with open lumen and with a greater number of endothelial cells [75]. Consequently, blood flow is sluggish and highly irregular.

Moreover, the neovessels exhibit a high resistance to capillary blood flow, and a low resistance to transcapillary flow hyperpermeability, resulting in a net efflux of fluid into the surrounding interstitial space where a lack of functional lymphatics allows it to accumulate, distending the elastic extracellular matrix and increasing the interstitial pressure [43]. As a consequence, tumour vessels exhibit chaotic arrangement and are compressed by the tumour cells, resulting in impaired microvascular perfusion which, in turn, promotes hypoxia and acidity in the tumour environment. This microenvironment, encompassing tumour-associated immunological tissue, is remodelled by the mast cells accumulated around the prostate cancer, resulting in higher cell density in the inner neoplastic tissue. Therefore, this process produces an increase in tissue elasticity and stiffness [76].

Metastasis and risk assessment

During the formation of a primary lesion, tumour cells undergo a variety of molecular events that eventually permit them to escape from the primary tumour site. First of all, prostate cancer, similarly to other forms of cancer, releases cells that invade the surrounding tissue and enter the circulation. These disseminated tumour cells circulate through the bloodstream, invade the lymph nodes, and usually localise in the bones, where they develop into metastatic lesions. The most common reported sites with metastatic involvement, after lymph node (95%) and bone metastasis (90%), are lung (46%), liver (25%), and adrenal (13%) [77]. There are several features of tumour vascularity that are associated with metastatic risk [78–80]. These include:

  • 1
    molecular-specific characteristics (e.g. VEGF);
  • 2
    spatial heterogeneity and chaotic structure;
  • 3
    fragile vessels with high permeability to macromolecules;
  • 4
    arteriovenous shunting and high vascular tortuosity;
  • 5
    intermittent or unstable flow;
  • 6
    uneven perfusion;
  • 7
    microvascular density;
  • 8
    increased cellular density.

These characteristics are detected by currently available, clinical imaging techniques.


  1. Top of page
  2. Abstract

Available imaging technology

Conventional CT, MRI, and TRUS imaging, have not proven sufficient sensitivity for timely and accurate prostate cancer diagnosis and prognosis. More recently, based on a better understanding of prostate cancer formation and growth, these conventional morphology based prostate imaging methods are being complemented by functional, metabolic, and molecular imaging techniques; one of the main objectives is the detection of angiogenesis [80,81]. In this review, only those imaging techniques aiming at the detection of tumour angiogenesis are discussed. Therefore, imaging methods, e.g. elastography and HistoScanning are not considered.


The MR signal is determined by hydrogen proton density and their inherently different longitudinal (T1) and transverse (T2) relaxation times [83]. T1-weighted image (T1-WI) sequences are not appropriate for prostate cancer detection, while conventional T2-weighted imaging (T2-WI) has shown good sensitivity but poor specificity in detecting prostate cancer (Table 1) [41,79,81,84–108]. More accurate prostate cancer detection requires a combination with metabolic and functional imaging. The most extensively used MRI techniques, combined with T2-WI, are dynamic contrast-enhanced MRI (DCE-MRI), diffusion-weighted imaging (DWI) and MR spectroscopic imaging (MRSI). Combination of these techniques is often referred to as multimodal MRI.

Table 1. Comparisons of imaging techniques for detection of prostate cancer
Imaging methodSensitivity, %Specificity, %Accuracy, %References
 Grey-scale39–7540–8250–87 [92,93,99]
 Colour Doppler27–7838–8040–59 [93,94,99]
 Power Doppler27–9335–7940–80 [94,95,100]
 Contrast-enhanced50–9041–9657–93 [96–98,101–103]
 Conventional T237–9627–6169–82 [41,81,104,105]
 DCE-MRI59–9674–9672–89 [79,84,106,107]
 DWI74–8557–9586–89 [84,85]
 MRSI57–9257–8867–87 [81,108]
 FDG4–7510083.3 [88,89]
 11C-choline33–10043–10071–93 [86,87]
 ProstaScint37–940–8664–86 [86,90,91]

DCE-MRI is performed after the i.v. injection of contrast agents. The most common contrast media are low-molecular-weight agents, e.g. gadolinium chelates, which rapidly diffuse in the extravascular extracellular space. After injection, analysis of tissue enhancement can yield information on blood perfusion and vascular wall permeability (Fig. 2a). The values of the above mentioned parameters are significantly different and specifically greater in malignant areas of tissue than in normal ones [81].


Figure 2. (a) Example of DWI-MRI image showing a suspicious cancerous area (indicated by T) with low signal intensity and (b) corresponding DCE-MRI image showing strong enhancement in the same area. Courtesy of Prof. Dr J. Barentsz, UMC St Radboud, Nijmegen, the Netherlands.

Download figure to PowerPoint

The enhanced MR signal in DCE-MRI can be interpreted by model fitting, e.g. by using the Tofts model for extravascular leakage [109]. The accuracy of DCE–MRI is reported in Table 1. The combination of high-spatial-resolution DCE-MRI (0.95 × 0.63 × 3 mm) and T2-WI yields improved assessment of extracapsular extension and better results for prostate cancer staging compared with either technique independently [110]. DCE-MRI can be performed on standard 1.5 T clinical systems using an endorectal coil and on 3 T systems without an endorectal coil [41]. Discrimination between BPH and prostate cancer in the transitional zone remains difficult. In addition, dynamic sequences are subject to the trade-off between spatial resolution (0.7–2 mm) and temporal resolution (2–25 s) [9,79].


Due to the higher cellular density of prostate cancer, the intracellular/extracellular volume ratio increases and water diffusion in the gross tumour volume is more restricted. Therefore, tumours appear with DWI as an area of decreased signal (Fig. 2b). Furthermore, DWI added to T2-WI has been proven to be better than T2-WI alone for detecting seminal vesicle invasion and for predicting recurrent cancer after radiation therapy. Combined DWI and T2-WI have also shown to be more specific but less sensitive than DCE-MRI in detecting cancer progression [84]. The advantages of DWI reside in the short acquisition time (could be <1.5 min), no need for administration of contrast agents or additional hardware, and high contrast between cancerous and normal tissues. However this technique is limited by poor spatial resolution (≈2 mm) and the potential risk of image distortion caused by post-biopsy haemorrhage [41,85].


MRSI provides metabolic information by assessing the relative concentration of cellular chemicals. The areas under the spectral peaks of the chemicals are related to their concentration, and changes in the concentrations of specific metabolites can be used to identify cancer with high specificity. In particular, prostate cancer shows lower levels of citrate and higher levels of choline relative to normal tissue. Based on this, MRSI calculates the ratio of choline and creatine to citrate [111]. The creatine peak is added because it is very close to the choline peak and, therefore, difficult to discriminate.

Scheidler et al.[82] showed that specificity of up to 91% is achieved when MRSI imaging is added to conventional MRI.

In general, MRSI is accepted to be useful for the evaluation of tumour location (both peripheral zone and transitional zone), local extent, volume, and aggressiveness. However, its main disadvantages are the poor spatial resolution (≈7 mm and 5 mm for 1.5 T and 3 T MRI scanner, respectively) and the long acquisition time (≈15 min) [112].

Nuclear imaging

Due to a high metabolic rate, prostate cancer consumes glucose through the glycolytic pathway, which is associated with higher glucose uptake. Nuclear imaging uses a radiolabeled analogue of glucose to identify cancerous lesions based on their increased metabolism. The distribution of the tracer, administered i.v., is imaged after ≈30 min by γ-cameras in single-photon emission computerised tomography (SPECT) and positron emission tomography (PET) [113].

Clinical PET systems use 4- to 6-mm detector pixels [114] and allow for an acquisition of 1–3 min per bed position [115]. Time-of-flight PET scans are currently under development with the aim of achieving time resolutions up to hundreds of picoseconds [115, L. Cosentino, unpublished data]. Recently introduced dual-modality PET/CT allows combining anatomical (CT) and metabolic (PET) information [86,87].

By far the most common PET radiotracer is fluorine-18 fluorodeoxyglucose (FDG), which images and quantifies the glucose turnover. Due to FDG limitations [88], new radiopharmaceuticals have been developed for imaging glucose metabolism, hypoxia, cellular proliferation, tumour receptors, angiogenesis, and gene expression. The most common PET radioisotopes are 18F-acetate, 11C-methionine, 18F-choline, 11C-acetate, and 11C-choline with promising results in imaging prostate cancer metastases [89,116].

Recently, 111 In-capromab pendetide (ProstaScint®) SPECT has been introduced as a promising option for detecting local recurrence or metastatic disease in prostate cancer [90]. ProstaScint is a specific agent that binds to an intracellular component of PSMA. After over one decade of experience, the value of ProstaScint imaging remains controversial [86,90,91].


Grey-level two-dimensional and three-dimensional TRUS is the most widely used method for prostate visualisation, volume measurement, and biopsy guidance [92,117]. However, grey-level TRUS is not adequate to provide accurate prostate cancer localisation [92,93].

With the aim of improving TRUS accuracy, new methods for TRUS angiogenesis imaging have been introduced; they mainly comprise Doppler ultrasonography (US) and contrast-enhanced US (CEUS) imaging [118,119].

Colour/power Doppler US

US colour Doppler quantifies blood velocity by estimating the frequency shift in the signal reflected from flowing blood cells. Areas with increased perfusion, as expected for prostate cancer, can therefore be identified. Power Doppler integrates the amplitude of the entire Doppler spectrum, without accounting for value and sign of the frequency shift. As a result, it is more sensitive than colour Doppler to perfusion, but it does not provide flow direction and quantification.

Colour and power Doppler offer, compared with grey-scale harmonic US, poorer temporal (25–50 Hz vs 200–500 Hz), and spatial (few mm vs 1 mm) resolution [78,94].

Although power Doppler is more sensitive than colour Doppler to low flow in small capillaries, it still insufficient to detect flow in angiogenic microvessels (10–50 µm diameter) (Table 1) [93–95].

More recently, the introduction of US contrast agents (UCAs) has provided new opportunities to boost the sensitivity of Doppler methods.


UCAs are micro-sized bubbles (microbubbles) made of gas encapsulated in a biocompatible elastic shell. Due to their size (1–10 µm in diameter) comparable with that of blood cells, they can pass the smallest microvessels [120].

Once administrated i.v., UCAs remain intravascular with a lifetime of several minutes. When invested by US waves, UCA microbubbles start oscillating, backscattering a large fraction of the received energy and therefore being easily detected by US imaging (Fig. 3). Up to single microbubbles can be detected, enabling imaging at the micrometre scale despite the resolution limits (millimetre scale) of US imaging. As a result, CEUS permits investigating the microvascular perfusion.


Figure 3. CEUS imaging of the prostate visualised at different times after an i.v. injection of an UCA bolus: (A) appearance time, (B) peak intensity, (C) wash out.

Download figure to PowerPoint

The behaviour of microbubbles differs depending on the acoustic intensity [118,120].

For very low energy (mechanical index <0.02), bubble compression and expansion are specular (linear behaviour) [121]. Under these conditions, the bubbles act like additional reflectors into the bloodstream thus increasing the sensitivity of colour and power Doppler, particularly in areas of increased vascular density, e.g. prostate cancer [34].

For intermediate energy levels (mechanical index 0.02–0.4) bubble expansion and compression are no longer specular [121]. This nonlinear behaviour gives rise to several strategies aimed at enhancing the signal reflected by the UCAs (nonlinear) while suppressing the signal reflected by tissue (linear) [121]. Increasing the contrast-to-tissue ratio in the reconstructed images results in increased sensitivity to UCAs (Table 1) and improved imaging of microvascular perfusion in cancer lesions [96,97].

At higher intensities (mechanical index >0.4) the microbubbles can be destroyed [122,123]. This feature is exploited by the so called flash-replenishment method [124] and the more recent intermittent-imaging method [125].

In general, possibly due to the approximations in the adopted models, up until now none of these perfusion quantification methods have shown ability to reliably localise prostate cancer [119,125,126].

Next to the UCA infusion methods, several more practical methods have been proposed that make use of a simple bolus injection. Analysis of the measured time–intensity curves permits the extraction of curve features, such as the area under the curve, peak intensity, or time to peak, which are related to perfusion [127]. More recently, a new local measurement has been proposed that aims at characterising the microvascular architecture by investigating the dispersion kinetics of UCAs through the microvasculature [98,128]. The preliminary results are promising [97] (Fig. 4) [128], but more extensive validation is still on-going.


Figure 4. Parametric dispersion map of the prostate based on the method in [128].

Download figure to PowerPoint

New developments of CEUS involve molecular imaging using site-targeted microbubbles, designed to bind to prostate cancer angiogenic expressions. Recently, Tardy et al.[129] have designed BR55, a VEGFR-2 specific UCA, showing that this agent not only provides information on tissue perfusion during the early vascular phase, but also it highlights the sites of active angiogenesis in rat prostate tumour. Research is on-going to confirm its potential for clinical application in human prostate cancer [130].


  1. Top of page
  2. Abstract

Pathological angiogenesis is a hallmark of cancer, being a fundamental part of a multistep process in the development of cancer progression, invasion, and metastasis.

Multiple structural and functional abnormalities of tumour blood vessels, e.g. altered perfusion, increased tortuosity, cellular proliferation, and microvascular density, together with heterogeneous microcirculation, dictate a hostile metabolic microenvironment characterised by hypoxia, acidosis, and glucose deprivation.

Angiogenesis is the key process in the development of tumour vasculature also for prostate cancer; every step is regulated by inducers (e.g. VEGF, TGF, MMPs) and inhibitors (e.g. angiostatin, endostatin, PSA) of angiogenesis. Recent advances in the understanding of vascular physiology and angiogenic mechanisms have led to the development of more sensitive approaches for prostate cancer diagnosis.

Functional CT is a high spatial resolution technique for assessing tumour neovasculature, but signal to noise ratio remains poor as compared with MRI.

MRI techniques are sensitive to various processes including blood flow, microvessel permeability and diameter, water diffusion, tissue oxygenation and metabolism. The complex interpretation of multimodal images [10], the impractical intraoperative use and high costs limit the use of MR methods at present.

PET/SPECT imaging can be used to directly evaluate various parameters related to tumour neovasculature, including direct evaluation of haemodynamic parameters (blood flow, blood volume), tissue properties (glucose metabolism, hypoxia) or expression of specific markers of angiogenesis (e.g. VEGF). However, this method is limited by high cost, short life of employed radioisotopes, and poor spatial resolution. TRUS, especially when combined with UCAs, can provide a practical option for prostate cancer detection. The injection of US microbubbles allows for improved detection of smaller and low-flow vessels. Currently, US and MRI remain the most widely studied and their clinical role in the management of prostate cancer is rapidly developing. However, these imaging techniques are not yet adopted for prostate interventions (biopsy targeting and focal therapy), as their accuracy is still questionable [9,10].

Molecular imaging offers additional opportunities for assessing prostate cancer angiogenesis through targeting of key molecules involved in the development and maintenance of tumour angiogenesis. It is necessary to identify new angiogenesis-related targets and optimise currently available imaging probes.

US imaging with molecularly targeted contrast microbubbles is of particular interest with this respect, as this method is quite inexpensive, does not involve ionising irradiation, and is widely available.

Given the challenges posed by accurate angiogenesis imaging and interpretation, it is likely that not one single parameter, target structure or imaging technique will enable it, but rather a fusion of multimodal images, which will allow for evaluation of the angiogenic cascade.

Considering the essential role of angiogenesis in tumour growth and spread, the prognostic potential of angiogenic activity measurement as well as the development of therapies that target angiogenesis, holds great promise. Anti-vascular or anti-angiogenic therapy might be a new treatment option to eradicate or at least control prostate cancer.


  1. Top of page
  2. Abstract