Clinical problem solving: utility of sonography in oncologic patients.

Pelvic sonography is used as the first-line modality for assessment of abnormal uterine bleeding. In the oncologic setting, a common indication for pelvic sonography arises in the breast cancer patient taking tamoxifen who presents with vaginal bleeding. Sonography remains the preferred modality for evaluation of those patients taking tamoxifen who have vaginal bleeding because of its anatomic detail, availability, safety profile, and cost-effectiveness. 39 Tamoxifen is a selective estrogen receptor modulator, commonly used in the adjuvant treatment of breast cancer. It has been shown to prevent breast cancer in high-risk women. 40 Although tamoxifen has antiestrogenic effects on breast tissue, it has a positive estrogenic effect on the postmenopausal endometrium and myometrium and results in an increased number of uterine and endometrial abnormalities. 41 , 42 A prospective study of 228 post-menopausal patients taking tamoxifen who were followed over 5 years revealed a statistically significant increased risk of endometrial lesions, including glandular hyperplasia and endometrial polyps. 41 In addition, findings from the National Surgical Adjuvant Breast and Bowel Project B-14 demonstrated a 2-fold relative risk of developing endometrial cancers in 2843 patients with estrogen receptor–positive invasive breast cancer treated with tamoxifen. 43 Approximately half of the women taking tamoxifen developed an endometrial lesion within 6 to 36 months of initiating treatment. 44 However, the National Surgical Adjuvant Breast and Bowel Project also demonstrated that the 5-year disease-free survival rate from breast cancer was 38% higher in women taking tamoxifen than in the placebo group, suggesting that the survival benefit provided by tamoxifen in women with breast cancer outweighed the smaller risk of endometrial cancer. 43 Given this increased risk of endometrial cancer, the breast cancer patient taking tamoxifen who presents with vaginal bleeding requires further assessment of the endometrium to assess for both benign and malignant disease related to tamoxifen use. Benign disease associated with tamoxifen use includes endometrial hyperplasia with or without cystic changes, endometrial polyps, adenomyosis, and growth of uterine fibroids. 45 Endometrial hyperplasia results from stimulation of the endometrial lining by the unopposed estrogenic effects of tamoxifen. Fishman et al 46 found that endometrial thickness increased with an increasing duration of tamoxifen use at a rate of 0.75 mm/y. After 5 years of use, the mean endometrial thickness was 12 mm (range, 6–21 mm). The endometrial thickness decreased by 1.27 mm/y after tamoxifen therapy was discontinued. Sonographic findings of endometrial hyperplasia include well-defined endometrial thickening with or without cystic changes. Less commonly, endometrial hyperplasia may be seen as more focal or asymmetric thickening with irregular margins. Sonohysterography may assist in identifying the diffuse nature of the endometrial thickening and guide management toward the more cost-effective blind endometrial biopsy, which could be performed in the outpatient setting. A second common endometrial lesion related to tamoxifen use is the endometrial polyp. The appearance of endometrial polyps also may vary on transvaginal sonography. They may appear as nonspecific endometrial thickening or as hyperechoic homogeneous well-defined lesions with smooth margins ( Figure 10 ). These lesions are best delineated by sonohysterography as smooth echogenic intracavity masses outlined by hypoechoic fluid. Doppler sonography may be helpful in identifying the polypoid stalk and point of attachment to the endometrial canal. Occasionally, polyps may contain cystic changes and appear heterogeneous ( Figure 11 ). Tamoxifen-related polyps are larger than those found in the general population, measuring up to 5 cm in mean diameter, and have been found to have a higher rate of malignant transformation. 47 , 48 Transvaginal sonography is used as an initial tool to identify a possible endometrial polyp, with further assessment by sonohysterography if needed to localize the lesion and guide subsequent management and resection. A third benign entity associated with tamoxifen use is adenomyosis or adenomyosis-like changes of the myometrium. Adenomyosis is histologically characterized by endometrial glands within the myometrium. Features of adenomyosis on sonography include areas of decreased echogenicity or heterogeneity of the myometrium, scattered myometrial cysts, and poor definition of the endomyometrial junction ( Figure 12 ). Real-time sonography for diagnosis of adenomyosis is essential, as the sonographic signs of adenomyosis are best identified during real-time scanning and are not reliably identified on static images. This entity is important, as it may mimic a falsely thickened endometrium on sonography in patients taking tamoxifen. Sonography remains a useful modality for identifying benign uterine conditions and guiding subsequent management, particularly in patients taking tamoxifen. Malignant disease secondary to tamoxifen use includes the development of endometrial carcinoma and uterine sarcomas. The endometrial cancer risk increases with both the duration of therapy and the cumulative tamoxifen dose. 47 Presently, there is no consensus regarding surveillance for endometrial cancer in patients receiving tamoxifen therapy. The American College of Obstetricians and Gynecologists and the Society of Obstetricians and Gynaecologists of Canada do not recommend screening by trans-vaginal sonography or endometrial biopsy for asymptomatic patients taking tamoxifen. 49 – 51 Premenopausal patients taking tamoxifen have no known increased risk of uterine cancer and require no additional monitoring. 40 , 51 Additional data have suggested that postmenopausal patients taking tamoxifen may be stratified into low- and high-risk groups before initiation of tamoxifen therapy. Berlière et al 52 evaluated 575 women with estrogen-positive breast cancer by transvaginal sonography before initiation of tamoxifen. Those women with an endometrial thickness of greater than 5 mm underwent hysteroscopy to evaluate for polyps or other endometrial lesions. Women found to have endometrial lesions before initiation of tamoxifen were at a higher risk of atypical endometrial lesions at the 2-year follow-up. 52 The American College of Obstetricians and Gynecologists recommends that postmenopausal women taking tamoxifen be closely monitored for symptoms of atypical endometrial hyperplasia or carcinoma and are encouraged to report abnormal vaginal bleeding. There is currently no consensus regarding a threshold endometrial measurement in asymptomatic women taking tamoxifen that should prompt biopsy. Asymptomatic women taking tamoxifen may have endometrial thickening measuring up to 9 to 13 mm. Many studies have demonstrated that screening using a threshold endometrial measurement for endometrial cancer in asymptomatic women taking tamoxifen is not effective. Fung et al 53 evaluated 304 women taking tamoxifen over a 6-year period and performed transvaginal sonography at study entry and then yearly. Endometrial biopsy was performed at study entry and repeated if substantial sonographic abnormalities were found or abnormal vaginal bleeding occurred. Fung et al 53 found that 32% of sonographic examinations revealed substantial endometrial abnormalities, with subsequent biopsies demonstrating that most (80%) represented benign polyps. An endometrial thickness of 9 mm had only 63% sensitivity and 60% sensitivity for detection of major uterine abnormalities and a positive predictive value of 1.4% for detecting endometrial cancer. Given the low specificity and positive predictive value, routine sonographic surveillance in asymptomatic women is not useful. Although the usefulness of screening pelvic sonography for breast cancer patients taking tamoxifen is controversial, sonography is helpful in the assessment of patients with abnormal uterine bleeding. 39 , 43 Endometrial cancer may appear as well-defined endometrial thickening or alternatively may be poorly defined, heterogeneous, and irregular on sonography. As imaging features of uterine malignancy can overlap those of benign uterine disease, biopsy is usually required for diagnosis. Additionally, more than one pathologic process may be present in the endometrium of patients taking tamoxifen, with coexistent polyps, hyperplasia, and carcinoma found concurrently at biopsy ( Figure 10 ).

Sonography is the initial modality for assessment of patients who present with suspected DVT. Venous sonography is the most accurate noninvasive test for the diagnosis of symptomatic proximal DVT. 1 Real-time assessment of vessel compressibility as well as evaluation of the venous waveforms and flow velocities allows identification of a thrombus within the vessel being interrogated. In addition, sonography remains readily available, easy to perform, and noninvasive for those patients with suspected DVT. Oncologic patients are at a 6-fold increased risk for the development of venous thromboembolism (VTE) compared to nononcologic patients, including both DVT and pulmonary embolism (PE). 2 This increased risk has been identified in both solid and hematologic malignancies and is highest among hospitalized cancer patients and those undergoing systemic or surgical treatment. 2 – 4 Other risk factors for development of VTE include patients receiving active systemic chemotherapy and those receiving erythropoiesis-stimulating agents as well as red blood cell and platelet transfusions. 5 – 7 Cancer patients who develop DVT are at more than 3 times greater risk of death than patients without cancer who develop DVT as well as patients with cancer but without DVT. 8 Patients with indwelling central venous catheters have been found to have a higher rate of upper extremity DVT, thought to be related to mechanical injury to the venous endothelium. 8 In a retrospective analysis of 573 patients, Giess et al 9 found that upper extremity venous thrombosis was common (40%) in symptomatic cancer patients and occurred twice as frequently in those cancer patients with indwelling central venous catheters. One prospective study investigating the incidence of catheter-related DVT in 92 patients with metastatic colorectal cancer showed that 73% of patients had a catheter-related thrombus, and 11% had obstruction of venous flow requiring anticoagulation. 8 The outcomes for patients with upper extremity DVT vary according to the presence or absence of cancer. 10 The Registro Informatizado de Enfermedad Tromboembólica registry, an ongoing registry of patients with confirmed symptomatic acute DVT or PE, looked at patients with upper extremity DVT to assess for the development of VTE and subsequent complications. It was found that catheter-related DVT was significantly more common in patients with cancer (53% versus 39%). 10 Additionally, 196 patients with both cancer and upper extremity DVT were found to have a worse prognosis than those patients without cancer, including a 9.7% incidence of recurrent DVT, symptomatic PE, or major bleeding complications. This group of patients was also found to have an increased risk of death when compared to those patients with upper extremity DVT but without cancer (22 versus 3.5%). 10 Given its high rate of complications and worse outcomes, upper extremity DVT in cancer patients, often a complication of central venous catheter placement, is frequently treated with anticoagulation therapy. Oncologic patients may present with symptoms such as upper extremity or facial swelling. Other times, routine restaging CT shows an apparent vascular filling defect within the upper extremity or neck vasculature. It is often unclear from the CT images whether this defect represents true vascular thrombosis or mixing of contrast material with unopacified blood ( Figures 1 and 2 ). Sonography is often used in these cases for clarity. On sonography, an acute thrombus appears as hypoechoic material partially or completely filling the lumen on grayscale imaging with an absence of color flow on Doppler imaging and absence of compressibility. A thrombus may be identified surrounding the echogenic linear central venous catheter. Many times, an apparent vascular filling defect identified on CT is indeed an artifact, resulting from inhomogeneous mixing of intravenous contrast material with unopacified blood ( Figure 2 ). Other times, no DVT will be identified, but real-time sonography documents the presence of slow flow within a vessel ( Figure 3 ). Careful assessment of the more central vasculature may identify the presence of stenosis or occlusion resulting in upstream slow flow. In addition to the assessment of DVT, sonography may be used as an adjunct in patients with suspected PE who cannot undergo the usual imaging examinations for PE such as pulmonary CT angiography because of poor renal function or a severe contrast agent allergy. Lower extremity sonography may also be useful in those patients with an indeterminate ventilation-perfusion scan. In these patients, determination of the presence or absence of VTE is extremely important, since cancer patients have a higher risk of developing PE compared to the general population. 2 Shinagare et al 11 studied 13,783 oncologic outpatients, including 395 patients with PE, and reported that central nervous system, lung/pleural, upper gastrointestinal, and pancreatic malignancies had a higher risk of PE compared to other types of malignancy. Given a high index of clinical suspicion, a lower extremity venous Doppler study is of particular value in these patients because positive study findings establish the diagnosis of VTE and prompt initiation of anticoagulation therapy. However, it should be noted that negative lower extremity Doppler study findings are not as useful in patients for whom there is a high clinical suspicion of PE and those with thromboembolic risk factors such as malignancy; if possible, these patients should be evaluated with CT angiography for more definitive assessment. 12

Another common indication for sonography of the oncologic patient is for evaluation of newly elevated liver function test results. Oncologic patients may have recently started new chemotherapeutic regimens and many have a known tumor burden within the liver. Common causes of acutely elevated liver function test results in the oncologic patient include steatosis secondary to chemotherapeutic agents, drug-induced acute hepatitis or cholecystitis, new bile duct obstruction, or new or worsening hepatic metastatic disease. Additionally, sonographic evaluation of the portal vein for the presence of a bland or malignant thrombus should be routinely performed in the oncologic patient. Finally, certain subtypes of oncologic patients may be at risk for hepatic veno-occlusive disease. In the general population, the prevalence of nonalcoholic fatty liver disease ranges from 3% to 24% and is more commonly associated with obesity, diabetes, men, and older age. 13 Sonography often plays an important role in the initial evaluation of elevated liver function test results. The presence of hepatic steatosis is commonly identified on many imaging modalities, including sonography, CT, and magnetic resonance imaging. Although sonography is often not the most reliable imaging tool for detecting hepatic steatosis, it gives a high degree of certainty of the presence of steatosis when positively identified on the sonogram. 14 A prospective study of 98 patients with elevated aminotransferase levels and increased liver echogenicity on sonography who underwent percutaneous biopsy showed steatosis of at least a moderate degree in 86%. 15 Oncologic patients undergoing chemotherapy are at an especially high risk for the development of hepatic steatosis. The liver plays a central role in metabolism, with many chemotherapeutic agents readily taken up by the liver. Up to 85% of cancer patients undergoing chemotherapy develop hepatic steatosis. 16 Sonography is often used for initial assessment of the oncologic patient with elevated liver function test results to evaluate for the presence of hepatic steatosis. In addition, sonography helps exclude other causes of elevated liver function test results, which will be discussed in subsequent sections. Many chemotherapeutic agents result in fatty infiltration of the liver, ranging from hepatic steatosis to steatohepatitis and, in extreme cases, hepatic fibrosis. Hepatic steatosis is commonly identified in patients undergoing treatment for breast and colorectal cancer. Examples of chemotherapeutic agents frequently associated with fatty infiltration include tamoxifen, irinotecan, 5-fluorouracil, and leucovorin. 17 – 19 In a retrospective study of 67 breast cancer patients treated with tamoxifen, 43% of the patients developed hepatic steatosis. 18 Steatosis has been identified in patients treated with molecular-targeted therapies including vascular endothelial growth factor inhibitors such as bevacizumab and tyrosine kinase inhibitors such as sunitinib. 20 Sonography shows increased echogenicity of the liver parenchyma, which may be diffuse, involving the whole liver, or have a well-defined geographic pattern. Diffuse fatty infiltration may result in increased attenuation of the ultrasound beam as it passes through the liver, causing posterior attenuation of the deep liver parenchyma and the diaphragm. Focal fatty deposition or focal sparing characteristically occurs in specific locations along the falciform ligament, along the gallbladder fossa, and within the porta hepatis. Regions of focal fatty sparing or areas of focal hepatic steatosis may be confused with an intrahepatic mass; however, the characteristic locations, well-defined geographic margins, and presence of unaltered blood vessels within the lesion identified during real-time sonography aid in diagnosis. Often times, hepatic steatosis may aid in identification of hepatic metastases. Metastases that are not well seen on CT because of the presence of marked diffuse hepatic steatosis are easily identifiable on sonography as hypoechoic masses in a background of diffuse increased echogenicity. Discontinuation of the offending drug may result in complete resolution of the fatty infiltration. Steatosis may progress to steatohepatitis, resulting in inflammation, cell injury, and fibrosis. This condition has been associated with a decreased hepatic reserve and ability to regenerate, resulting in increased 90-day postoperative mortality in certain patient groups with body mass indices above 25 kg/m 2 . 17 , 21 Prompt identification and reporting of hepatic steatosis in the oncologic patient are important because they may change patient treatment. Sonography is the initial modality for assessment of possible acute cholecystitis given that it is rapid, accurate, and noninvasive. It allows evaluation of the gallbladder wall thickness and hyperemia as well as the presence of gallstones, sludge, and pericholecystic fluid. In addition, real-time sonography allows for assessment of the sonographic Murphy sign, a highly specific sign for the presence of acute cholecystitis. 22 In the oncologic patient, many chemotherapies have been associated with cholecystitis, with early recognition and rapid access to sonography important for diagnosis. Drug-induced cholecystitis is a well-known complication of many chemotherapeutic agents. Many prior reports in the literature have described the well-known complication of drug-induced acute or chronic cholecystitis as a result of hepatic arterial infusion of chemotherapeutic agents. 23 , 24 High concentrations of chemotherapy are delivered directly to the liver by the hepatic artery, resulting in gallbladder toxicity, likely by direct perfusion of the cystic or accessory cystic artery. The reported incidence of cholecystitis secondary to hepatic arterial infusion in the literature is around 1% to 2%, although the incidence of gallbladder infarction ranges from 1% to 41%. 24 This complication is now frequently avoided by prophylactic cholecystectomy before the start of hepatic arterial infusion therapy. Many other systemic chemotherapies have also been associated with the development of cholecystitis. A retrospective review of 54 pediatric oncology patients with various malignancies undergoing chemotherapy demonstrated the development of acalculous cholecystitis in 6% of patients. 25 Although a rare complication, drug-induced acute cholecystitis has also been associated with several newer oncologic agents, including tyrosine kinase inhibitors as well as mammalian target of rapamycin inhibitors such as everolimus ( Figure 4 ), 26 – 28 although the incidence is not yet well described. Treatment includes broad-spectrum antibiotics and withholding the offending chemotherapeutic agent, usually with resolution of symptoms and radiologic findings. Other patients may require a cholecystostomy tube, cholecystectomy, or both. Sonography is often the initial noninvasive imaging study for assessment of biliary obstruction. Real-time sonography provides the opportunity for careful assessment of the biliary tree and has been advocated as first-line imaging for evaluation of biliary obstruction given that it is cost-effective, readily available, and noninvasive. In addition, sonography has good accuracy for delineating the level of biliary obstruction. A retrospective analysis of 42 patients who presented with biliary ductal dilatation on sonography revealed that sonography could delineate the level of obstruction in 88% and identify the cause of obstruction in 48%. 29 Many cancer patients present with rising liver function test results during the course of therapy and undergo sonographic assessment to evaluate for potential biliary obstruction. Biliary ductal obstruction may occur at the level of the common bile duct secondary to extrinsic compression from adjacent lymphadenopathy ( Figure 5 ) or bulky masses. Alternatively, it may be due to primary or secondary tumors of the biliary tree. Malignant strictures may occur from a hepatobiliary malignancy such as cholangiocarcinoma or may be secondary to metastatic disease to the liver and biliary tree that extends along the common duct and central intrahepatic biliary tree ( Figure 6 ). The presence of biliary ductal dilatation and rising liver function test results as well as comparison with prior imaging studies could help confirm the presence of new or acute obstruction. Careful assessment of the liver and biliary tree with sonography can identify the presence of biliary obstruction and often the level and cause of obstruction, which aids in management and may preclude the need for other imaging. 30 Portal vein thrombosis occurs in many settings, including hepatic cirrhosis, primary and secondary malignancies of the liver, hypercoagulable states, and infectious or inflammatory processes of the liver, pancreas, or bowel. Malignant portal vein thrombosis is a complication of primary hepatic malignancy or secondary to metastatic disease, commonly seen with hepatocellular carcinoma and pancreatic cancer. Sonography plays an essential role in assessment of patients suspected of having portal vein thrombosis, for both initial diagnosis as well as distinguishing bland from malignant portal vein thrombosis. Sonographic assessment of the portal vasculature in the oncologic patient is helpful to assess for patency, the direction of flow, and the presence of a thrombus. If a thrombus is detected, sonography may be helpful in distinguishing between a bland and tumor thrombus. A tumor thrombus on sonography is seen as an expanded portal vein containing echogenic material within the lumen, often with abnormal vascularity within ( Figure 7 ). The presence of arterial Doppler signals within the thrombus has moderate sensitivity and high specificity for diagnosing malignant portal vein thrombosis. 31 , 32 Dodd et al 31 performed color and pulsed Doppler sonography in 47 patients with cirrhosis and portal vein thrombosis, with detection of pulsatile flow in the portal vein thrombus yielding 62% sensitivity and 95% specificity for diagnosis of malignant portal vein thrombosis. More recently, contrast-enhanced sonography has been shown to be the most sensitive and specific sonographic technique. 33 Chronic portal vein thrombosis may result in cavernous transformation of the portal vein ( Figure 7 ), seen as multiple serpiginous vessels in the porta hepatis surrounding the thrombosed or stenosed portal vein. Distinguishing a tumor thrombus from a bland thrombus is essential, as the presence of a tumor thrombus worsens the prognosis substantially and may preclude patients from undergoing additional medical or surgical interventions. Hepatic veno-occlusive disease, also known as sinusoidal obstruction syndrome, is a well-recognized complication after hematopoietic stem cell transplantation. It is a serious complication of stem cell transplantation that may occur secondary to cytoreductive chemotherapy or radiation therapy from stem cell transplantation, most frequently encountered in the first 20 days after transplantation. 34 Although the diagnosis is often suspected clinically using formal criteria such as the Baltimore and Seattle criteria, sonography remains the first-line imaging investigation for further supporting the diagnosis. 35 Grayscale sonography identifies key features of hepatic veno-occlusive disease within the liver and gallbladder, and pulsed Doppler imaging allows assessment of the direction and waveform of the portal veins, which are not distinguished on other cross-sectional imaging modalities. Additionally, sonography can exclude the presence of malignant hepatic infiltration and other biliary diseases. The reported prevalence of hepatic veno-occlusive disease after transplantation varies but may occur in up to 54% of stem cell transplant recipients. 36 , 37 Alternatively, hepatic veno-occlusive disease may occur after intensive chemotherapy regimens in patients with hematologic malignancies, even in the absence of stem cell transplantation. The diagnosis may be suspected from the clinical picture of painful hepatomegaly, jaundice, ascites, or unexplained weight gain. Although the etiology is unclear, hepatic veno-occlusive disease is thought to be secondary to injury to the hepatic sinusoidal endothelium with subsequent occlusion of the hepatic venules. This condition leads to obstruction of hepatic sinusoidal outflow, hepatic congestion, and subsequent hepatic fibrosis. 34 The diagnosis may be challenging given that the imaging features are nonspecific, with many conditions presenting with similar clinical and imaging features, most notably graft-versus-host disease and infectious or drug-induced hepatitis. In equivocal cases, patients may require liver biopsy to distinguish hepatic veno-occlusive disease from other entities such as graft-versus-host disease, although liver biopsy has been associated with an increased rate of bleeding complications in this patient population due to coagulopathy and thrombocytopenia. 38 Characteristic sonographic features can support the diagnosis and may preclude the need for liver biopsy. Most commonly, sonographic findings of hepatic veno-occlusive disease include ascites and marked gallbladder wall thickening, often greater than 6 to 8 mm in thickness ( Figure 8 ). 34 Doppler sonography may show pulsatile flow in the portal vein, followed by decreased velocity, biphasic flow, or both. Ultimately, hepatofugal portal venous flow results ( Figure 8 ). 34 Although reversal of flow in the portal vein is thought to be a specific sign, it usually occurs late in the course of the disease. A recannalized umbilical vein, abdominal varices, and splenomegaly may also be seen. Currently, there is no specific treatment for hepatic veno-occlusive disease. Some patients may be treated with antithrombotic therapy and immunosuppression. Prevention and early diagnosis are essential to decrease the high morbidity and mortality related to this disease, with prompt sonographic evaluation important. In summary, sonography remains important in the global evaluation of the liver parenchyma, gallbladder, biliary tree, and hepatic vasculature in the oncologic patient for addressing causes of elevated liver function test results. Sonography often identifies the cause of the elevated liver function test results, as outlined above. Sometimes, no cause is identified for the patient’s elevated liver function test results. In these cases, other causes must be considered, including drug-induced hepatitis and an overall worsening of the hepatic tumor burden ( Figure 9 ).

This review has illustrated the unique advantages that sonography can offer in the oncologic setting over other imaging modalities, including real-time and focused assessment of specific patient symptoms and signs, morphologic and functional evaluation with grayscale and Doppler imaging during a single examination, absence of ionizing radiation, and increased cost-effectiveness over other radiologic imaging studies. In addition, sonography is readily available, easy to perform, and noninvasive. Although cross-sectional imaging has increasingly become the primary imaging modality for evaluation of the oncologic patient, sonography remains an extremely useful tool for evaluation of various conditions and clinical scenarios associated with cancer.