The development of a viable tumor blood supply is essential to support tumor growth and proliferation.² This may occur through sprouting from preexisting vessels, de novo vascular formation through recruitment of circulating endothelial progenitor cells, and vessel cooption. The tumor microenvironment including hypoxia, glucose deprivation, and low pH plays an important role in the initiation of tumor angiogenesis via the activation of oncogenes and/or inactivation of tumor suppressor genes.³

In terms of morphology, the tumor vasculature appears functionally distinct and spatially heterogeneous. The vasculature lacks the usual hierarchical branching pattern and in general demonstrates greater vessel density at the tumor periphery. The vessels themselves are thin and tortuous, characterized by a relatively high endothelial cell proliferation rate, incomplete endothelium, relative absence of smooth muscle or pericyte investiture, and hyperpermeability resulting in high interstitial fluid pressure.

Microvessel density (MVD) and vascular endothelial growth factor (VEGF) are commonly used immunohistologic measures of angiogenesis. In a number of cancers including lung, renal, gastrointestinal, and pancreatic cancer, associations have been found between MVD and/or VEGF and various perfusion CT parameters. Most of the evidence relates to lung cancer. Typically, these correlations have been moderate: histologic analysis has been varied, based on different numbers of “hotspot” counts or from random areas of the whole tumor. There have also been negative studies, in part reflecting the heterogeneity of analyses and immunohistologic biomarkers used.⁴

With respect to the lung, peak CT enhancement in patients with solitary pulmonary nodules has been correlated significantly with both MVD and VEGF, irrespective of the benign or malignant nature of the nodules.⁵ In patients with operable non-small cell lung cancer CT peak enhancement, blood flow, blood volume or permeability surface area product have been shown to have a moderate correlation with MVD: Li et al. found that the CT regional blood flow correlated with CD34 expression (r = 0.715; p ≤ 0.001) assessed in six tumor regions: central (three regions) and peripheral (three regions).⁶ Ma et al. have shown that CT peak enhancement and regional blood flow correlate with CD34 expression assessed in five hotspots in VEGF-positive but not VEGF-negative tumors.⁷ Similarly, Sauter et al. have found a moderate correlation between extraction fraction and blood flow and CD34,⁸ while Spira et al. have reported a positive association between MVD and blood flow and volume.⁹ Peak enhancement, blood flow, and relative blood volume have also been shown to be significantly higher in VEGFpositive compared to VEGF-negative tumors.⁷,¹⁰

There have also been several pathologic correlative studies in abdominally sited cancers. In renal cell cancer an initial study in 24 patients showed a moderate correlation (r = 0.60) between peak enhancement and hotspot MVD (CD34).¹¹ More recently, a further small study (n = 10) where patients with renal cell cancer underwent volumetric perfusion CT prior to surgery has confirmed that regional blood flow and blood volume correlated significantly with MVD (CD34; r = 0.600 to 0.829); however, extraction fraction (measured as the transfer constant, K^trans) only demonstrated moderate correlations with MVD in nonnecrotic areas (r = 0.550).¹²

Moderate correlations (r = 0.42) between regional blood volume and hotspot MVD (CD34) have been found in gastric adenocarcinoma.¹³ Similarly in colorectal cancer, moderate correlations between regional blood volume and permeability surface area product and nonhotspot MVD (CD34) have been shown.¹⁴ In pancreatic adenocarcinoma, moderate correlations (r = 0.49) have also been found between MVD and peak enhancement in the arterial phase.¹⁵

Perfusion CT parameters may also inform on the presence of perfusion-related hypoxia. In lung cancer, blood volume measurements have been found to be negatively correlated to an exogenous marker of hypoxia (pimonidazole: r = -0.48).¹⁶ However, one of the challenges of clinicopathologic correlative studies is the comparison of in vivo with ex vivo findings. Tacelli et al. have showed that the areas of low regional tumor blood volume but high permeability have higher CD34 expression (assessed in three hotspots in the nonnecrotic tumor portion) than areas of high regional tumor blood volume and high permeability: 72.1 versus 47.9, p = 0.038.¹⁷ They postulated this difference could be related to the effect of the tumor microenvironment.¹⁷

The dose and manner in which contrast agents are administered may influence parameter quantification. Kinetic modeling benefits from a bolus injection of contrast agent, at an injection rate of at least 4 mL/s via a large bore intravenous cannula, usually sited in the antecubital fossa. Injection rates beyond 10 mL/s appear to confer no additional benefit for quantification. The total injected iodine dose should be within the range of 12 to 18 g. It is important that the iodine concentration administered is not less than 300 mg/mL. Conversely, if iodine concentrations are greater than 350 mg/mL the contrast agent must be warmed to body temperature prior to injection, as the higher viscosity will slow the actual injection rate. The volume of contrast agent will depend on the iodine concentration used.¹

With current state-of-the-art technology, an entire organ may be encompassed with a high spatial resolution and a high temporal acquisition sampling rate. To ensure accurate quantification, a sampling rate of 2 seconds or less should be used (Fig. 32-4). For the initial perfusion phase, an acquisition duration of 45 seconds is adequate; for the interstitial phase at least five additional time points are recommended, the sampling rate ranging from 5 to 15 seconds depending on the kinetic model applied.¹ As the concentration of iodine within blood vessels and tissues is proportional to the resultant increase in attenuation, temporal changes in attenuation can be analyzed using standard kinetic models without prior conversion to iodine concentration.

For target areas located in motion susceptible sites, breath-holding during the first pass study will reduce motion artifact and voxel displacement. For longer-duration studies, motion correction/image registration may compensate for motion artifact, and this is incorporated into commercial software platforms, for example, based on a nonrigid deformable registration technique.¹⁸

The vascular parameters that can be quantified include regional blood flow, blood volume, and extraction fraction or permeability surface area product (Table 32.1). Different mathematical models may be applied to derive these. Commonly used tracer kinetic modeling includes “Maximum initial slope”, “Patlak” and “Distributed parameter analysis” (Table 32.2). In practice, quantification using commercial platforms requires a region of interest to be placed within an input artery (to generate an arterial input function [AIF]) and for the lesion(s) of interest. From the arterial and tissue enhancement-time curves that are displayed by the software, quantitative parametric maps are generated for each pixel (or voxel) representing a parameter value (Fig. 32-5). While volume-ofinterest analysis, encompassing the whole tumor, may be the least susceptible to observer error and experience, and provide a global evaluation of perfusion and angiogenesis, this may not best reflect the heterogeneity within a tumor, particularly if this demonstrates areas of necrosis, calcification or hemorrhage, as these are “averaged” in the process.¹⁹

In practice, it is important to choose an appropriate target lesion(s). Ideally, this should be greater than 2 cm and not in close proximity to large vessels (e.g., superior vena cava or aorta), heart or diaphragm in order to reduce artifacts. For lesions located in motion susceptible sites, breath-holding techniques (+/- initial hyperventilation), and/or the use of motion correction/image registration will reduce motion artifacts. For target lesions located within the bowel, the use of a standard antiperistaltic (Buscopan or Glucagon) will reduce the motion related to peristalsis. It is also important to select an appropriate artery to derive the AIF. The vessel chosen must be of sufficient size to prevent partial volume averaging occurring secondarily due to pulsation or movement artifact (including peristalsis) or laminar flow within the vessel. If this is the case this will influence and change the arterial time-attenuation curve. It is assumed that the arterial time-attenuation curve of the artery in the field of view is identical to the vessel supplying the tumor. This is only valid if there is no significant stenosis of the artery or other major feeding vessels.

Quantitative parameters derived from perfusion CT may be used to monitor the effects of a variety of treatments. These include chemotherapy with standard and novel agents (antiangiogenic drugs, vascular disrupting agents, and immunotherapy), radiotherapy, and interventional oncologic procedures such as embolization of the tumor vascular supply or radiofrequency ablation. Early evidence has shown that a common long-term effect of treatment is a reduction in perfusion CT parameters following treatment completion, although in the short-term there may be a variable vascular effect related to the therapeutic mechanism of action (Table 32.3).

With standard chemotherapy, which affects actively replicating cells via DNA damage or interruption of DNA repair, this effect is thought to reflect the loss of angiogenic cytokine support following cell death.⁵² With antiangiogenic therapies, differing vascular effects may be seen depending on the mechanism of action of the drug under investigation and timing of the scan. An initial effect may be a decrease in vascular permeability and reduction in interstitial fluid pressure, with normalization of function of the vasculature resulting in a transient increase in tumor blood flow.⁵³ In the longer term, with subsequent pruning of the vasculature, a reduction in regional blood flow, blood volume, and vascular permeability may be elicited (Fig. 32-6). With vascular disrupting agents, which target the proliferating immature vasculature +/- the mature vasculature a rapid shutdown in tumor vascularization may occur that is usually transient and reversible within 24 to 48 hours. This may be followed by a rebound revascularization.⁵⁴ With radiotherapy, the acute effects are related to an initial inflammatory effect; the permeability is related to microvascular damage, which can lead to tumor shrinkage.⁵⁵ With interventional procedures perfusion CT parameters may provide evidence of effective treatment or the need for further procedural attempts for optimal therapeutic effect.⁵⁶