Contrast ultrasonography: Cancer characterization (Proceedings)

Grey-scale ultrasound has proven to be modestly sensitive for neoplastic nodules in the liver. In cases where there are many nodules, especially larger or coalescing nodules, ultrasound is more likely to detect the nodules. When the nodules are isoechoic, few in number, very small or located in difficult to image portions of the liver, the nodules may go undetected. Even when the nodules are seen, the confidence regarding whether they are benign "old dog" changes or malignancy may be quite low. This presentation will review the current state-of-the-art in ultrasound imaging for the detection and characterization of liver nodules.

Contrast media

Ultrasound contrast agents are basically bubbles of encapsulated gas. The capsule constituents of commercial contrast agents vary with the manufacturer, but most commonly is an inert lipoprotein in 2nd generation agents. The shell of Definity® is composed of a mixture of 3 lipids having molecular weights ranging from 670 to 5750. Only one contrast agent, Optison®, with a shell composed of human albumin, may be immune-active. Due to reported allergic reactions, Optison® is considered contraindicated for use in dogs and cats.

The luminal gas also varies between manufacturers. In first generation contrast agents such as Levovist®, air was the gas used. But air is not a powerful sound generator and been replaced in 2nd generation agents with inert gases. Examples of these inert gases in the newer 2nd generation agents include octofluoropropane (Definity®) and sulfur hexafluoride (SonoVue®).

These encapsulated bubbles need to be small enough to pass through the lungs without significant attrition and, therefore, reach the liver in sufficient numbers to provide enhancement. Bubbles made by agitation of saline are too large (>50 µm) to pass through the lungs, are very short-lived and are not as reflective as commercial contrast agents. Unlike the chemical contrast agents of CT and MR, ultrasound contrast agents have a fairly wide distribution of sizes, depending on type and manufacturer. Definity® has a spectrum of size ranging from 1.1 to 3.3 µm with 98% less than 10 µm. At this size, large numbers of reach the abdominal structures after 1st pass through the lung.

Contrast software

Imaging these bubbles is a technological challenge. Many ultrasound innovations have been developed as a corollary of contrast imaging. Tissue harmonic ultrasound technology was developed incidental to contrast ultrasound research. The goal of contrast ultrasound technology is to elicit a large sound signal, larger than from the tissues, from these bubbles. Until recently this was accomplished through harmonic technology. Bubbles are much more powerful emitters of harmonic signal (=nonlinear response) than any biological tissue. Each bubble, depending on type and size, has a characteristic resonant frequency. At this resonant frequency the bubbles emit a disproportional large amount of harmonic signal. The trick is to insonify the bubbles with the correct magnitude (mechanical index) of resonant frequency sound and then filter all the returning signal to isolate the bubble portion.

Harmonic ultrasound has been the most common method that takes advantage of the nonlinear response of bubble contrast agents to their respective resonant frequency. Harmonic frequency are multiples of the transmitted sound frequency (4 MHz transmitted has a harmonic frequency of 8 MHz). Various technologies exists to eliminate the fundamental transmitted signal and isolate the harmonic bubble signal. Many of these methods suffer from a low signal-to-noise ratio, or are difficult to optimize on the particular machine system. The harmonic signal is always much weaker the transmitted fundamental.

A newer technology involves the isolation of the bubble signal from the nonlinear fundamental component. This involves utilization of phase and amplitude changes unique to the nonlinear bubble response. The advantage to this technique is the comparative high signal-to-noise ratio of the fundamental signal (compared to a harmonic component) and ability to image at higher frequencies.

Liver nodules in human patients

Worldwide, ultrasound is the most widely available imaging modality for screening for liver pathology. Routine grey-scale ultrasound falls well short of contrast-enhanced CT and MR for accuracy of liver nodule detection and characterization. In Europe and the far east, where the technologies are approved for clinical usage, contrast ultrasound has shown great improvement compared to routine ultrasound and comparable results to CT and MRI for liver nodule characterization.

Hepatocellular carcinoma (HCC) is the most common primary liver tumor in human patients. On tri-phasic contrast-enhanced CT and MRI the detection rate for all HCC masses is reported to be from 89-92%, but considerably lower for nodules smaller than 2 cm.> Contrast-enhance ultrasound has improved sensitivity for lesion detection from 45-55% with grey-scale ultrasound to 80-90%, with reference images obtained with spiral CT imaging. The largest gain was in detection of smaller nodules. The typical HCC lesion has high and rapid peak enhancement, quick wash-out phase (during normal liver portal enhancement), chaotic peritumoral and intralesional tortuous vessels and vascular lakes. The most suitable timing is during the arterial phase due to visualization of enhanced blood flow.

Metastatic lesion detection also benefits from contrast-enhancement. Most lesions, regardless of primary malignancy type, are seen as hypoechoic nodules (black holes) compared to a hyperechoic (whiter) background of normal liver enhancement. After contrast enhancement the lesions were more conspicuous compared to normal liver and 10-94% more lesions were identified in 68% of the patients compared to grey-scale ultrasound. Approximately equal numbers were identified only by contrast ultrasound and helical CT.

The premise of improved contrast-enhanced conspicuity of malignant nodules for CT and contrast-enhanced ultrasound is a presumption of primarily arterial vessels induced during malignant neovascularization of hepatic masses. A primarily arterial blood supply would allow for an earlier peak wash-in phase, compared to the delayed peak associated with the predominant portal blood supply to normal liver.

Liver nodules in dogs

Recent studies have confirmed the safety and efficacy of contrast ultrasound for imaging the liver of normal dogs and no side effects have yet been reported for Definity® or SonoVue® in dogs or cats. Peak enhancement times in normal livers ranged from 22 (±7) to 46 (±19) seconds. The time to peak was measured differently in the two studies. Normal liver enhancement in dogs and cats is uniform and coincidental to portal blood flow timing. Separation of hepatic arterial and portal phases of the inflow pattern has not yet been possible.

In a case series of 32 dogs with liver nodules, seventeen dogs had benign nodules and 15 dogs had malignant nodules. The sensitivity (100%), specificity (94.1%), positive predictive value (93.8%), negative predictive value (100%) and overall accuracy (96.9%) for the characterization of the nodules was highly significant (P<0.0001). Most benign nodules, regardless of their initial echogenicity, had enhancement patterns similar to the normal surrounding liver and appeared isoechoic (= "disappeared") at the time of peak liver enhancement. Only one nodule, a hepatoma, had perfusion characteristics similar to malignancy. All malignant nodules were hypoechoic, or had large hypoechoic portions, during the peak portal phase concurrent with normal liver enhancement. Tortuous afferent vessels were noted in the periphery of HCC lesions during the arterial phase. Most carcinoma and lymphoma nodules had rapid uniform wash-in during the arterial phase, rapid wash-out during the portal phase (= early wash-in, early wash-out"), and appeared as hypoechoic nodules (black holes) compared to the uniformly peak-enhanced normal surrounding liver. Hemangiosarcoma nodules never contrast enhanced, had tortuous peripheral afferent vessels and were detectable throughout normal liver enhancement as hypoechoic lesions (black holes).

A more recent study indicated a increased ability of contrast-enhanced ultrasound to detect hemangiosarcoma nodules. Three dogs with large splenic masses and normal liver on routine ultrasound examination had small nodules (5-8 mm diameter) seen on contrast ultrasound. The primary mass and liver nodules were confirmed as hemangiosarcoma on surgical excision and histopathology.

Splenic nodules

Very little may be extrapolated from human ultrasound imaging regarding splenic lesions. Hemangiosarcoma is not a common primary malignancy and the spleen is an uncommon site of primary or metastatic neoplasia

A recent abstract review the contrast ultrasound characteristics of 17 dogs and 2 cats with splenic nodules. All benign lesions were isoechoic to the spleen at peak enhancement, except the hematoma which was uniformly poorly enhanced. Lymphoma (3) and malignant histiocytosis (1) lesions had shorter time to peak enhancement (early wash-in) and wash-out compared to the surrounding normal spleen. A fine net of thin uniformly distributed vessels was present in the lymphoma nodules. Hemangiosarcomas (3) and undifferentiated sarcoma (1) had homogeneous anechoic areas with highly vascularized surrounding parenchyma.

Renal nodules

The renal perfusion pattern in normal dogs is biphasic. The initial fast intense inflow with prolonged peak in the cortex is followed by a much more gradual inflow to the medullary region. Peak perfusion in the cortex in normal dogs is approximately 13 (± 5) seconds and 21 (±6) seconds in the medullary region.

Limited information is available on the appearance of renal masses. Because of the structure of the cortex and medulla, detection of nodules can be difficult. Anecdotal evidence indicates that sarcomas are poorly perfused in dogs, as previously indicated in the liver and spleen.

Detection of acute tumor infarction is not possible in grey scale ultrasound. Chronic non-neoplastic infarctions have a classic appearance with increased echogenicity and flattening of the outer contour, both characteristics indicating a benign condition with advanced chronicity and likely reperfusion of the region. Neoplastic-related renal infarctions are readily detected with contrast ultrasound and characterized by regional, usually wedge-shaped, hypoechoic cortical perfusion deficits. The incidence of subclinical infarction associated with metastatic neoplasia is not well understood. This may represent embolic disease related to subclinical DIC or neoplastic emboli. Increased coagulability has been reported as a paraneoplastic syndrome.

Future considerations

One of the more exciting developments in contrast ultrasound is molecular imaging. With molecular imaging compounds can be attached to the contrast bubble. Commercially available contrast agents and kits to affix biotinylated proteins, including antibodies and other proteins, are currently available for research use. Molecular imaging is a noninvasive approach to determine the expression of indicative marker molecules of the tumor angiogenesis process. Meanwhile, this approach has been established for all imaging modalities and may further improve sensitivity of diagnostic tumor imaging. Another goal is to provide information with respect to drug treatment monitoring and therapeutic vascular targeting strategies. Research indications for these agents include characterization of tumor angiogenesis, myocardial infarction, thrombus and ischemic-reperfusion models. The agents appear useful in mouse, rat, and dog models of disease and await further evaluation for veterinary clinical applications, especially in the area of cancer research.

Summary

Contrast ultrasound has clinical application for improved detection and characterization of liver, spleen and kidney nodular lesions in dogs. Clinical or research implementation of this imaging modality requires familiarity with contrast agents and advanced ultrasound technology. An important research corollary to contrast ultrasound is molecular imaging which allows bubbles to be directed to specific binding sites by attaching the contrast agent monoclonal antibodies or other marker proteins for functional imaging of cancer- and treatment-related marker molecules.

References

O'Brien RT. Improved detection of metastatic hepatic hemangiosarcoma nodules with contrast ultrasound in three dogs. Vet Radiol Ultrasound 2007; 48(2):146-8.

O'Brien RT, Iani M, Matheson J, Delaney F, Young K. Contrast harmonic ultrasound of spontaneous liver nodules in 32 dogs. Vet Radiol Ultrasound 2004;45(6):547-53.

Ohlerth S, O'Brien RT. Contrast ultrasound: general principles and veterinary clinical applications. Vet J 2007;174(3):501-12.

Rossi F, Leone VF, Vignoli M, Laddaga E, Terragni R. Use of contrast-enhanced ultrasound for characterization of focal splenic lesions. Vet Radiol Ultrasound 2008; 49(2):154-64.