3304jum557-712online_Layout 1 3/19/14 11:30 AM Page 575
ORIGINAL RESEARCH
Double–Contrast-Enhanced Sonography for Diagnosis of Rectal Lesions With Pathologic Correlation Man Lu, MD, Brain Yan, MD, Jun Song, MD, Wu Ping, MD, Lin-Xian Yue, MD, Bin Song, MD Objectives—Transabdominal sonography with a gastrointestinal contrast agent has been widely used in China for investigation of digestive disorders. Double–contrastenhanced sonography combines a gastrointestinal luminal contrast agent with an intravenous contrast agent for imaging of lesions. The purpose of this pilot study was to assess the value of double–contrast-enhanced sonography for preoperative diagnosis of rectal lesions. Methods—We conducted a prospective single-center study using double–contrastenhanced sonography of rectal lesions. Patients were administered both rectal and intravenous contrast agents, and imaging was performed transabdominally, transanally, and transrectally. Morphologic characteristics and perfusion parameters were compared between histologically proven adenocarcinomas, adenomas, and inflammatory masses. Perfusion parameters were analyzed with time-intensity curves, measuring the contrast arrival time, time to peak, peak intensity, and area under the curve of the lesions and normal rectal tissue.
Received April 11, 2013, from the Department of Ultrasound, Sichuan Academy of Medical Sciences and Sichuan Province People’s Hospital, Chengdu, China (M.L., J.S., W.P., L.-X.Y.); Division of Gastroenterology, Department of Medicine, Western University, London, Ontario, Canada (B.Y.); and Department of Radiology, West China Hospital of Sichuan University, Chengdu, China (B.S.). Revision requested May 3, 2013. Revised manuscript accepted for publication August 2, 2013. This work was supported by the Sichuan Province Science and Technology Fund. Address correspondence to Bin Song, MD, Department of Radiology, West China Hospital of Sichuan University, 37 Guo Xue Xiang, 610041 Chengdu, Sichuan, China. E-mail: [email protected] Abbreviations
AUC, area under the curve; ROI, region of interest doi:10.7863/ultra.33.4.575
Results—From January 2009 to September 2012, 420 patients were recruited, with 227 patients meeting inclusion/exclusion criteria and having 232 rectal lesions analyzed (172 rectal adenocarcinomas, 45 adenomas, and 15 inflammatory masses). Adenocarcinomas had variable enhancement patterns. Adenomas were all hypoenhanced in a homogeneous pattern. Inflammatory masses had a hyperenhanced rim with no central enhancement. Time-intensity curve perfusion parameters (arrival time, time to peak, peak intensity, and area under the curve) of rectal adenocarcinomas, adenomas, and inflammatory masses were significantly different compared to normal rectal tissue (P< .05). The differences in the arrival time, peak intensity, and time to peak among the different lesions were also significant (P < .05). Conclusions—Double–contrast-enhanced sonographic assessment of morphologic enhancement patterns combined with vascularity parameters may help differentiate benign and malignant rectal lesions. Key Words—contrast-enhanced sonography; gastrointestinal ultrasound; rectal lesion; time-intensity curve parameters
R
ectal cancer is one of the most common malignancies worldwide.1,2 Early detection and stage-appropriate treatment are critical for long-term survival. Indeed, the estimated 5-year survival has improved in the past years to 43%.1–4 In the pathogenesis, premalignant intraepithelial neoplasia located in a rectal adenoma precedes the occurrence of invasive rectal cancer.5 Accurate preoperative diagnosis and staging of rectal lesions are essential for treatment planning.6 The choice of therapy is dependent on the type of lesion, location, size, depth of invasion, and presence of lymph node or distant metastases.
©2014 by the American Institute of Ultrasound in Medicine | J Ultrasound Med 2014; 33:575–583 | 0278-4297 | www.aium.org
3304jum557-712online_Layout 1 3/19/14 11:30 AM Page 576
Lu et al—Double–Contrast-Enhanced Sonography for Diagnosis of Rectal Lesions
Currently, transrectal sonography is a highly sensitive technique that provides highly accurate assessment of the tumor penetration depth. This technique has been shown to be a reliable tool for the staging of rectal cancer.7 However, to optimize the focal distance between the probe and rectal wall, a balloon on the probe is used to distend the rectum (by injecting 20–30 mL of water). Unfortunately, the lesions can be flattened by the rectal balloon, resulting in imaging artifacts and inaccuracies in staging. Because of the disturbing effects of gastrointestinal luminal contents, conventional sonography does not easily visualize gastric or intestinal tumors. The likelihood of detecting gastric or intestinal cancers is improved by using water or a gastrointestinal contrast agent, which can reliably evaluate the bowel wall structure and improve the detection of wall changes compared to conventional sonography.8,9 More recently, the use of intravenous contrastenhanced sonography offers a unique opportunity to image real-time tissue perfusion with a blood pool tracer and is increasingly being used to characterize the vasculature of lesions or an organ of interest. By showing exquisitely detailed vascularity and tissue perfusion in real time with excellent spatial resolution, contrast-enhanced sonography can assess tumor angiogenesis in vivo.10,11 Kurihara et al12 demonstrated that the perfusion patterns seen with contrast-enhanced sonography reflect the vascular structure of tumors, therefore helping differentiate benign from malignant conditions. Published guidelines have thus included the use of transabdominal contrast-enhanced sonography for the workup of liver, pancreatic, and kidney lesions.13 Currently, contrast-enhanced sonography has been applied to staging and directing therapeutic procedures to the myocardium, liver, kidney, pancreas, skin, prostate, breast, and lymph nodes even in the presence of tissue motion.11–13 To our knowledge, the role of contrast-enhanced sonography in rectal lesions has not been evaluated. Transabdominal sonography with a gastrointestinal intraluminal contrast agent has been increasingly used in China to detect digestive disorders.8,9 Double–contrastenhanced sonography combines a gastrointestinal luminal contrast agent with an intravascular contrast agent for imaging of lesions.8,9 This technique can optimize morphologic assessment by providing a good acoustic window and allows vascularity assessment to differentiate benign from malignant lesions. In this study, we used double– contrast-enhanced sonography to evaluate rectal lesions and compare the imaging and histologic findings. To our knowledge, this study was the first to assess double– contrast-enhanced sonographic features of rectal lesions with histologic results.
576
Materials and Methods Patient Population From January 2009 to September 2012, 420 consecutive patients who underwent double–contrast-enhanced sonography for rectal disorders were considered for enrollment in the study. All patients were required to have prior endoscopic confirmation of a rectal lesion and biopsy for inclusion in the study. The exclusion criteria were as follows: (1) no rectal lesion on endoscopy; (2) inability to tolerate the enema preparation or cooperate with the examination; (3) contraindications to intravenous ultrasound contrast agents; (4) receipt or upcoming receipt of neoadjuvant chemoradiation and radiotherapy; and (5) incomplete data or double–contrast-enhanced sonograms available for analysis. All patients with tumors underwent resection 1 week after the double–contrast-enhanced sonographic examination. This protocol was approved by the Institutional Review Board and Ethics Committee of our hospital. All patients provided written informed consent to participate in the study, which included consent for intravenous contrast agent administration. Double–Contrast-Enhanced Sonographic Protocol Equipment Double–contrast-enhanced sonographic examinations were performed with a MyLab 90 ultrasound system (Esaote SpA, Genoa, Italy) equipped with contrast-tuned imaging technology and Analysis QontraXt version 4.0 software. A transrectal biplanar probe (axial and sagittal, TRT33; linear bandwidth, 4–13 MHz; convex bandwidth, 9–3 MHz) and an abdominal probe (CA431; convex bandwidth, 1–8 MHz) were used for rectal imaging. A low–mechanical index, real-time, contrast-specific contrast-enhanced sonographic mode was used during the examination, with the mechanical index set between 0.1 and 0.2. Tianxia (Huzhou, China) is a luminal contrast agent supplied as a powder composed of a type of common Yan rhizome derivative (50 g per package). The powder was reconstituted with 700 mL of boiling water and gentle agitation to form a homogeneous suspension. The suspension was allowed to cool to 35°C to 40°C before instillation into the rectum through a disposable enemator. SonoVue (Bracco SpA, Milan, Italy) was used as the intravenous contrast agent. SonoVue is a second-generation blood pool contrast-enhancing agent consisting of microbubbles containing a sulfur hexafluoride gas stabilized by a phospholipid shell. J Ultrasound Med 2014; 33:575–583
3304jum557-712online_Layout 1 3/19/14 11:30 AM Page 577
Lu et al—Double–Contrast-Enhanced Sonography for Diagnosis of Rectal Lesions
Procedure Before imaging, the patient was placed in the supine position, and a digital rectal examination was performed to assess the approximate location and extent of the lesion. The rectal luminal contrast agent was then administered to displace air and achieve distension, resulting in the appearance of the lumen as a homogeneous acoustic window. The volume of the intraluminal contrast agent ranged from 500 to 1000 mL, depending on the distention condition of the rectum. After the rectal lumen was distended, a sonographic examination of the rectum was performed by transabdominal and transanal approaches with the patient in the supine position using the abdominal probe, and then transrectal sonography was conducted using the biplanar probe covered with a natural rubber latex cover with the patient in the left lateral decubitus position. When a tumor was found, the location, shape, and largest diameter of the lesion were recorded for later analysis. Under transrectal sonography with luminal contrast imaging, intravenous contrast-enhanced sonography of the lesion was performed in the “contrast general” mode after the administration of 2.4 mL of SonoVue (sulfur hexafluoride–based microbubbles at 8 μL [45 μg]/mL and 2.5 μm in diameter) as a bolus injection via a 19-gauge peripheral venous cannula, followed by a 10-mL saline flush. Real-time contrast imaging was recorded continuously for 6 minutes and stored on the hard drive of the system for documentation and analysis. All examinations were performed by a single physician (M.L.), who had 20 years of experience with abdominal sonography and 5 years of experience with contrast-enhanced sonography and was blinded to the histologic results.
Image Analyses The cine loops from double–contrast-enhanced sonography were reviewed retrospectively by 2 independent radiologists (J.S. and W.P.), who were experts in diagnostic sonography and contrast imaging and were also blinded to the histologic results. In cases of discrepancies, the reviewers assessed the saved images together and reevaluated their findings to reach a consensus. A normal rectal wall was defined as a 5-layer echo structure (Figure 1): The first hyperechoic and second hypoechoic layers correspond to the mucosa, the third hyperechoic layer to the submucosa, the fourth hypoechoic layer to the muscularis propria, and the fifth hyperechoic layer to the subserosa and serosa. The enhancement of the lesions was compared to the surrounding normal parenchyma of the rectal wall to determine the speed (fast or slow wash-in and wash-out phases) and pattern of enhancement. Enhancement patterns were characterized into 3 types: hyperenhanced, hypoenhanced, and unenhanced. The imaging results were compared with the histologic results. Contrast-enhanced time-intensity curves were used to assess the perfusion characteristics of the lesions by offline measurements with the Analysis QontraXt version 4.0 software. Regions of interest (ROIs) of similar size were drawn by each observer on the lesions and the normal circumrectal tissues carefully to avoid the lumen and areas of necrosis (appearing as contrast-filling defects). The timeintensity curve was obtained from the stored video clips after the ROIs were set. The ROI position was adjusted manually from frame to frame as necessary. The contrast arrival time, time to peak, peak intensity, sharpness, and area under the curve (AUC) within the rectal lesions and
Figure 1. Transrectal sonograms of a normal rectal wall visualized as a 5-layer echo structure (arrows). A, Before intraluminal contrast agent administration. B, After instillation of the contrast agent. The rectal lumen was dilated by the contrast agent, which can serve as a homogeneous acoustic window and can better delineate the wall of the rectum. A
J Ultrasound Med 2014; 33:575–583
B
577
3304jum557-712online_Layout 1 3/19/14 11:30 AM Page 578
Lu et al—Double–Contrast-Enhanced Sonography for Diagnosis of Rectal Lesions
the normal circumrectal tissues were computed by the built-in software. The parameter arrival time (in seconds) was defined as the first point of the curve clearly above the baseline intensity, followed by a rise. The time to peak (in seconds) was defined as the time from the start of the injection to the maximum intensity of the curve. The peak intensity (in decibels) was defined as the maximum intensity. The AUC can be directly estimated from the integral of the signal intensity versus time curve. Parameters obtained from 3 ROIs were averaged, and the values were taken as the representative parameters for a single time. A duplicate measurement was done by the same reviewer. Finally, the representative parameter values were the averages of the measurements by the 2 reviewers. Statistical Analyses Statistical analyses were performed with a standard statistical software package (SPSS version 18.0; IBM Corporation, Armonk, NY). Time-intensity curve perfusion parameters, including the arrival time, peak intensity, and AUC, were normally distributed continuous data and were expressed as mean ± standard deviation. Differences in time-intensity curve perfusion parameters between different lesions and normal perirectal tissue were compared by an independent samples t test. Comparisons of time-intensity curve parameters among the 3 rectal lesions and normal perirectal tissue were performed with 1-way analysis of variance, and a Student-Newman-Keuls test was conducted for pair-wise comparison of perfusion parameters among the different lesions and normal perirectal tissue. P < .05 was considered statistically significant.
Results Of 420 patients referred for double–contrast-enhanced sonography, 227 patients with 232 lesions (5 patients had 2 lesions) met the inclusion/exclusion criteria and were included in this study. There were 143 male and 84 female patients with a mean age of 60 ± 13 years. All 227 patients tolerated the examinations well with no immediate adverse events. The mean lesion size was 5.4 cm (range, 2.5–10 cm). Histopathologic examination showed 172 adenocarcinomas, 45 adenomas, and 15 inflammatory masses. Histopathologic stages of adenocarcinomas were 34 in T1, 56 in T2, 70 in T3, and 12 in T4. Eighty-one lesions were located in the lower third (0– 5 cm) of the rectum (7 inflammatory masses, 18 adenomas, and 56 adenocarcinomas), 86 in the middle third (5–8 cm; 4 inflammatory masses, 15 adenomas, and 67 adenocarcinomas), and 65 in the higher third (8–12 cm; 4
578
inflammatory masses, 12 adenomas, and 49 adenocarcinomas). There were 56 lymph node metastases among 114 total lymph nodes and 8 distant metastases. Double–contrast-enhanced sonography of adenocarcinomas showed a hyperenhanced homogeneous appearance in 43 of 172 lesions (25%; Figure 2), a hyperenhanced heterogeneous pattern in 112 of 172 (65%), and hypoenhanced and unenhanced patterns in 17 of 172 (10%). Seventeen of 45 adenomas showed a hypoenhanced homogeneous pattern, whereas 28 of 45 had a hyperenhanced homogeneous appearance on double– contrast-enhanced sonography (Figure 3). All 15 inflammatory lesions showed a hyperenhanced rim pattern without central enhancement (Figure 4). Table 1 shows comparisons between time-intensity curve perfusion parameters for different lesions and normal rectal tissue. Analysis of variance showed significant differences between different lesions (F = 87.72; P < .001). Adenocarcinomas had a faster arrival time and time to peak than normal rectal tissue, and the differences were significant (P < .05; Figure 2). Adenomas had a later time to peak and lower peak intensity than normal rectal tissue, and the differences were significant (P < .01; Figure 3). Inflammatory lesions had a faster arrival time and higher peak intensity than normal rectal tissue, and the differences were significant (P < .05; Figure 4). Statistically significant differences in perfusion parameters were seen between adenocarcinomas, adenomas, and inflammatory lesions (Table 1). Adenocarcinomas had a higher peak intensity and faster arrival time than adenomas, and the differences were significant (P < 0.001; Figure 5). Adenocarcinomas had a higher peak intensity and faster time to peak than inflammatory lesions, and the differences were significant (P < .01). Inflammatory masses had a faster arrival time, higher peak intensity, and later time to peak than adenomas, and the differences were significant (P < .05).
Discussion Colorectal cancer is the third most common cancer worldwide.1 Approximately 40% to 50% of colorectal cancers are located in the rectum. More than 90% of cases of rectal cancer develop from benign adenomas, but only approximately 10% of rectal adenomas will eventually develop into rectal cancer.7 Benign adenomas only require local excision with either endoscopic or surgical polypectomy techniques. For rectal lesions that are not amenable to endoscopic removal or in cases of uncertainty, accurate preoperative assessment of the lesion is important for coun-
J Ultrasound Med 2014; 33:575–583
3304jum557-712online_Layout 1 3/19/14 11:30 AM Page 579
Lu et al—Double–Contrast-Enhanced Sonography for Diagnosis of Rectal Lesions
seling the patient regarding the best options for removal, perioperative risks, and chance of recurrence.14,15 For a small but crucial group of patients, differentiation of a benign rectal mass from malignant neoplasia remains an unsolved problem despite improved technology. Furthermore, inflammatory masses are challenging to diagnose and differentiate from carcinomas. Standard sonography often is of little benefit, as inflammatory masses appear isoechoic or hypoechoic to the surrounding tissue, similar to rectal cancer. The difference is more easily distinguishable when the mass includes an abscess fluid pocket. To our knowledge, this study was the first to use a quantitative analysis of double–contrast-enhanced sonography for discrimination of rectal adenocarcinomas from
masslike adenomas and inflammatory masses. Chung et al16 demonstrated that a luminal contrast agent was a useful addition to sonographic examination of the bowel. It provided reliable evaluation of the bowel wall structure and improved the detection of wall changes compared to conventional sonography. In our study, we used a new luminal contrast agent to provide a homogeneous acoustic window. The luminal contrast agent provided excellent delineation of the 5-layer echo pattern of the normal rectal wall and the morphologic characteristics of rectal lesions. Villous rectal lesions appear as a polypoid or fat during endoscopy or standard transrectal sonography. In this study, 4 villous adenomas were observed and appeared to “float” within the contrast agent–filled lumen.
Figure 2. Images from a 63-year-old male patient with an adenocarcinoma in the middle third of the rectum. A, Endoscopic image showing the adenocarcinoma. B, Microscopic view of a hematoxylin-eosin–stained specimen showing the adenocarcinoma (original magnification ×100). C, After rectal instillation of the contrast agent, transrectal sonography with luminal extension showed a heterogeneous hypoechoic mass with irregular margins. D, After intravenous contrast agent injection, double–contrast-enhanced sonography in the harmonic imaging mode showed a hyperenhanced homogeneous adenocarcinoma with a clear margin and a time-intensity curve measurement indicating a low-perfusion pattern in the normal rectal wall. E, Regions of interest of the time-intensity curve were drawn within the lesion and a showed a high-perfusion pattern in the lesion.
J Ultrasound Med 2014; 33:575–583
579
3304jum557-712online_Layout 1 3/19/14 11:30 AM Page 580
Lu et al—Double–Contrast-Enhanced Sonography for Diagnosis of Rectal Lesions
Adenocarcinoma sonographically appears as an irregular and “fixed” or “immobile” lesion. In this study, double– contrast-enhanced sonography clearly demonstrated that even a very small adenocarcinoma (1 cm) appeared as a polypoid and fixed lesion within the contrast agent–filled cavity. Therefore, a gastrointestinal luminal contrast agent could improve the morphologic assessment of rectal lesions to help differentiate malignant from benign lesions. This study demonstrates that aggressive malignant tumors have a high degree of angiogenesis. In our experience, after intravenous administration of the contrast agent, adenocarcinomas displayed marked and heterogeneous signal enhancement in 90% of cases. However, as tumors increase in size, they may outgrow their blood supply, resulting in necrosis within the tumors.17,18 Larger tumors in our study tended to be hypoenhanced or unenhanced, indicating that the lesions had necrotic changes. These features were absent in inflammatory lesions. The enhancement pattern of the inflammatory lesions included a hyperenhanced rim appearance and no central enhancement. In this study, all adenomas had homogeneous enhancement, and most of the large adenomas had a hyperenhanced pattern. This preliminary study shows that
differences in enhancement patterns may help in the differentiation of rectal lesions. Double–contrast-enhanced sonography not only showed high-resolution imaging of the morphologic characteristics of rectal tumors but also provided useful, albeit more limited, quantitative parameters of tumor perfusion characteristics. Time-intensity curve parameters provide a quantitative means to differentiate benign from malignant rectal lesions. These data are easy to interpret and can be readily implemented into routine clinical practice. In this study, when compared to normal rectal tissue, we found that rectal adenocarcinomas had a faster arrival time and time to peak, adenomas had a lower peak intensity, and inflammatory lesions had an earlier arrival time and a higher peak intensity. Our study also showed that the vascular supply of rectal lesions differed significantly between benign and malignant lesions, which could be used as an adjunctive diagnostic tool in challenging cases. There were some limitations to this study. First, we did not study the relationship between the enhanced intensity revealed by double–contrast-enhanced sonography and microvessel density in rectal lesions. Further investi-
Figure 3. Images from a 47-year-old female patient with a rectal adenoma. A, Endoscopic image showing a tubulovillous adenoma with no malignant features. B, Microscopic view of a hematoxylin-eosin–stained specimen showing the tubulovillous adenoma (original magnification ×100). C, Transrectal sonography with luminal extension by the contrast agent showed a heterogeneous hypoechoic mass with regular margins in the lower third of the rectum. D, With intravenous contrast agent injection, double–contrast-enhanced sonography in the harmonic imaging mode showed a hyperenhanced homogeneous adenoma with a clear margin and a time-intensity curve measurement indicating a low-perfusion pattern in the lesion. TP indicates time to peak.
580
J Ultrasound Med 2014; 33:575–583
3304jum557-712online_Layout 1 3/19/14 11:30 AM Page 581
Lu et al—Double–Contrast-Enhanced Sonography for Diagnosis of Rectal Lesions
gation on this parameter may improve the diagnostic power of double–contrast-enhanced sonography. Second, the placement and depth of the ROI may have influenced the parameter results. We used the average of multiple different measurements from independent reviewers in an attempt to minimize the effect of ROI selections. Third, there is certain subjectivity to assessing morphologic enhancement, such as determining homogeneous versus heterogeneous patterns. Given that this modality is heavily operator dependant, accuracy may be reduced in nonexpert hands.
In conclusion, we have provided a preliminary report of evaluation and diagnosis of rectal lesions by using a new double–contrast-enhanced sonographic technique. Double–contrast-enhanced sonography can improve and overcome some limitations of traditional transrectal sonography by providing an optimal acoustic window for both grayscale and contrast imaging of the rectal lesions. Morphologic enhancement characteristics and quantitative perfusion parameters may help differentiate rectal adenocarcinomas from adenomas or inflammatory lesions. Thus, double–contrast-enhanced sonography may become a useful method for characterizing and evaluating the blood supply of rectal lesions to differentiate benign from malignant lesions.
Figure 4. Images from a 27-year-old male patient with an inflammatory mass in the lower third of the rectum. A, Endoscopic image showing the inflammatory lesion. B, Microscopic view of a hematoxylin-eosin–stained specimen showing the inflammatory lesion (original magnification ×100). C, Transrectal sonography with color Doppler showed a heterogeneous hypoechoic mass (calipers) with irregular margins in the lower third of the rectum. D, Double–contrast-enhanced sonography with intravenous contrast agent injection showed a hypoenhanced heterogeneous mass with a time-intensity curve measurement indicating very low perfusion inside the mass. E, The time-intensity curve measurement of the enhanced rim of the lesion showed relatively high perfusion compared to that of inside the lesion. TP indicates time to peak.
J Ultrasound Med 2014; 33:575–583
581
3304jum557-712online_Layout 1 3/19/14 11:30 AM Page 582
Lu et al—Double–Contrast-Enhanced Sonography for Diagnosis of Rectal Lesions
Table 1. Time-Intensity Curve Values for the Different Rectal Lesions Lesion
n
Adenocarcinoma 172 Adenoma 45 Inflammatory mass 15 Normal perirectal tissue 232 P (t test) Adenocarcinoma vs adenoma Adenocarcinoma vs inflammatory mass Adenocarcinoma vs normal perirectal tissue Adenoma vs inflammatory mass Adenoma vs normal perirectal tissue Inflammatory mass vs normal perirectal tissue
AT, s
TP, s
PI, dB
AUC
19.9 ± 3.5 24.4 ± 1.7 20.7 ± 3.4 24.1 ± 2.5
25.5 ± 77 25.2 ± 12.5 42.9 ± 26.1 41.4 ± 17.2
46.4 ± 9.2 16.5 ± 5.9 62.0 ± 8.9 44.3 ± 8.9
6.6 ± 3.9 4.0 ± 2.9 3.8 ± 4.8 4.2 ± 3.3
.99 .99 .013
>.99 .007 .035 .008 .035 >.99
ORIGINAL RESEARCH
Double–Contrast-Enhanced Sonography for Diagnosis of Rectal Lesions With Pathologic Correlation Man Lu, MD, Brain Yan, MD, Jun Song, MD, Wu Ping, MD, Lin-Xian Yue, MD, Bin Song, MD Objectives—Transabdominal sonography with a gastrointestinal contrast agent has been widely used in China for investigation of digestive disorders. Double–contrastenhanced sonography combines a gastrointestinal luminal contrast agent with an intravenous contrast agent for imaging of lesions. The purpose of this pilot study was to assess the value of double–contrast-enhanced sonography for preoperative diagnosis of rectal lesions. Methods—We conducted a prospective single-center study using double–contrastenhanced sonography of rectal lesions. Patients were administered both rectal and intravenous contrast agents, and imaging was performed transabdominally, transanally, and transrectally. Morphologic characteristics and perfusion parameters were compared between histologically proven adenocarcinomas, adenomas, and inflammatory masses. Perfusion parameters were analyzed with time-intensity curves, measuring the contrast arrival time, time to peak, peak intensity, and area under the curve of the lesions and normal rectal tissue.
Received April 11, 2013, from the Department of Ultrasound, Sichuan Academy of Medical Sciences and Sichuan Province People’s Hospital, Chengdu, China (M.L., J.S., W.P., L.-X.Y.); Division of Gastroenterology, Department of Medicine, Western University, London, Ontario, Canada (B.Y.); and Department of Radiology, West China Hospital of Sichuan University, Chengdu, China (B.S.). Revision requested May 3, 2013. Revised manuscript accepted for publication August 2, 2013. This work was supported by the Sichuan Province Science and Technology Fund. Address correspondence to Bin Song, MD, Department of Radiology, West China Hospital of Sichuan University, 37 Guo Xue Xiang, 610041 Chengdu, Sichuan, China. E-mail: [email protected] Abbreviations
AUC, area under the curve; ROI, region of interest doi:10.7863/ultra.33.4.575
Results—From January 2009 to September 2012, 420 patients were recruited, with 227 patients meeting inclusion/exclusion criteria and having 232 rectal lesions analyzed (172 rectal adenocarcinomas, 45 adenomas, and 15 inflammatory masses). Adenocarcinomas had variable enhancement patterns. Adenomas were all hypoenhanced in a homogeneous pattern. Inflammatory masses had a hyperenhanced rim with no central enhancement. Time-intensity curve perfusion parameters (arrival time, time to peak, peak intensity, and area under the curve) of rectal adenocarcinomas, adenomas, and inflammatory masses were significantly different compared to normal rectal tissue (P< .05). The differences in the arrival time, peak intensity, and time to peak among the different lesions were also significant (P < .05). Conclusions—Double–contrast-enhanced sonographic assessment of morphologic enhancement patterns combined with vascularity parameters may help differentiate benign and malignant rectal lesions. Key Words—contrast-enhanced sonography; gastrointestinal ultrasound; rectal lesion; time-intensity curve parameters
R
ectal cancer is one of the most common malignancies worldwide.1,2 Early detection and stage-appropriate treatment are critical for long-term survival. Indeed, the estimated 5-year survival has improved in the past years to 43%.1–4 In the pathogenesis, premalignant intraepithelial neoplasia located in a rectal adenoma precedes the occurrence of invasive rectal cancer.5 Accurate preoperative diagnosis and staging of rectal lesions are essential for treatment planning.6 The choice of therapy is dependent on the type of lesion, location, size, depth of invasion, and presence of lymph node or distant metastases.
©2014 by the American Institute of Ultrasound in Medicine | J Ultrasound Med 2014; 33:575–583 | 0278-4297 | www.aium.org
3304jum557-712online_Layout 1 3/19/14 11:30 AM Page 576
Lu et al—Double–Contrast-Enhanced Sonography for Diagnosis of Rectal Lesions
Currently, transrectal sonography is a highly sensitive technique that provides highly accurate assessment of the tumor penetration depth. This technique has been shown to be a reliable tool for the staging of rectal cancer.7 However, to optimize the focal distance between the probe and rectal wall, a balloon on the probe is used to distend the rectum (by injecting 20–30 mL of water). Unfortunately, the lesions can be flattened by the rectal balloon, resulting in imaging artifacts and inaccuracies in staging. Because of the disturbing effects of gastrointestinal luminal contents, conventional sonography does not easily visualize gastric or intestinal tumors. The likelihood of detecting gastric or intestinal cancers is improved by using water or a gastrointestinal contrast agent, which can reliably evaluate the bowel wall structure and improve the detection of wall changes compared to conventional sonography.8,9 More recently, the use of intravenous contrastenhanced sonography offers a unique opportunity to image real-time tissue perfusion with a blood pool tracer and is increasingly being used to characterize the vasculature of lesions or an organ of interest. By showing exquisitely detailed vascularity and tissue perfusion in real time with excellent spatial resolution, contrast-enhanced sonography can assess tumor angiogenesis in vivo.10,11 Kurihara et al12 demonstrated that the perfusion patterns seen with contrast-enhanced sonography reflect the vascular structure of tumors, therefore helping differentiate benign from malignant conditions. Published guidelines have thus included the use of transabdominal contrast-enhanced sonography for the workup of liver, pancreatic, and kidney lesions.13 Currently, contrast-enhanced sonography has been applied to staging and directing therapeutic procedures to the myocardium, liver, kidney, pancreas, skin, prostate, breast, and lymph nodes even in the presence of tissue motion.11–13 To our knowledge, the role of contrast-enhanced sonography in rectal lesions has not been evaluated. Transabdominal sonography with a gastrointestinal intraluminal contrast agent has been increasingly used in China to detect digestive disorders.8,9 Double–contrastenhanced sonography combines a gastrointestinal luminal contrast agent with an intravascular contrast agent for imaging of lesions.8,9 This technique can optimize morphologic assessment by providing a good acoustic window and allows vascularity assessment to differentiate benign from malignant lesions. In this study, we used double– contrast-enhanced sonography to evaluate rectal lesions and compare the imaging and histologic findings. To our knowledge, this study was the first to assess double– contrast-enhanced sonographic features of rectal lesions with histologic results.
576
Materials and Methods Patient Population From January 2009 to September 2012, 420 consecutive patients who underwent double–contrast-enhanced sonography for rectal disorders were considered for enrollment in the study. All patients were required to have prior endoscopic confirmation of a rectal lesion and biopsy for inclusion in the study. The exclusion criteria were as follows: (1) no rectal lesion on endoscopy; (2) inability to tolerate the enema preparation or cooperate with the examination; (3) contraindications to intravenous ultrasound contrast agents; (4) receipt or upcoming receipt of neoadjuvant chemoradiation and radiotherapy; and (5) incomplete data or double–contrast-enhanced sonograms available for analysis. All patients with tumors underwent resection 1 week after the double–contrast-enhanced sonographic examination. This protocol was approved by the Institutional Review Board and Ethics Committee of our hospital. All patients provided written informed consent to participate in the study, which included consent for intravenous contrast agent administration. Double–Contrast-Enhanced Sonographic Protocol Equipment Double–contrast-enhanced sonographic examinations were performed with a MyLab 90 ultrasound system (Esaote SpA, Genoa, Italy) equipped with contrast-tuned imaging technology and Analysis QontraXt version 4.0 software. A transrectal biplanar probe (axial and sagittal, TRT33; linear bandwidth, 4–13 MHz; convex bandwidth, 9–3 MHz) and an abdominal probe (CA431; convex bandwidth, 1–8 MHz) were used for rectal imaging. A low–mechanical index, real-time, contrast-specific contrast-enhanced sonographic mode was used during the examination, with the mechanical index set between 0.1 and 0.2. Tianxia (Huzhou, China) is a luminal contrast agent supplied as a powder composed of a type of common Yan rhizome derivative (50 g per package). The powder was reconstituted with 700 mL of boiling water and gentle agitation to form a homogeneous suspension. The suspension was allowed to cool to 35°C to 40°C before instillation into the rectum through a disposable enemator. SonoVue (Bracco SpA, Milan, Italy) was used as the intravenous contrast agent. SonoVue is a second-generation blood pool contrast-enhancing agent consisting of microbubbles containing a sulfur hexafluoride gas stabilized by a phospholipid shell. J Ultrasound Med 2014; 33:575–583
3304jum557-712online_Layout 1 3/19/14 11:30 AM Page 577
Lu et al—Double–Contrast-Enhanced Sonography for Diagnosis of Rectal Lesions
Procedure Before imaging, the patient was placed in the supine position, and a digital rectal examination was performed to assess the approximate location and extent of the lesion. The rectal luminal contrast agent was then administered to displace air and achieve distension, resulting in the appearance of the lumen as a homogeneous acoustic window. The volume of the intraluminal contrast agent ranged from 500 to 1000 mL, depending on the distention condition of the rectum. After the rectal lumen was distended, a sonographic examination of the rectum was performed by transabdominal and transanal approaches with the patient in the supine position using the abdominal probe, and then transrectal sonography was conducted using the biplanar probe covered with a natural rubber latex cover with the patient in the left lateral decubitus position. When a tumor was found, the location, shape, and largest diameter of the lesion were recorded for later analysis. Under transrectal sonography with luminal contrast imaging, intravenous contrast-enhanced sonography of the lesion was performed in the “contrast general” mode after the administration of 2.4 mL of SonoVue (sulfur hexafluoride–based microbubbles at 8 μL [45 μg]/mL and 2.5 μm in diameter) as a bolus injection via a 19-gauge peripheral venous cannula, followed by a 10-mL saline flush. Real-time contrast imaging was recorded continuously for 6 minutes and stored on the hard drive of the system for documentation and analysis. All examinations were performed by a single physician (M.L.), who had 20 years of experience with abdominal sonography and 5 years of experience with contrast-enhanced sonography and was blinded to the histologic results.
Image Analyses The cine loops from double–contrast-enhanced sonography were reviewed retrospectively by 2 independent radiologists (J.S. and W.P.), who were experts in diagnostic sonography and contrast imaging and were also blinded to the histologic results. In cases of discrepancies, the reviewers assessed the saved images together and reevaluated their findings to reach a consensus. A normal rectal wall was defined as a 5-layer echo structure (Figure 1): The first hyperechoic and second hypoechoic layers correspond to the mucosa, the third hyperechoic layer to the submucosa, the fourth hypoechoic layer to the muscularis propria, and the fifth hyperechoic layer to the subserosa and serosa. The enhancement of the lesions was compared to the surrounding normal parenchyma of the rectal wall to determine the speed (fast or slow wash-in and wash-out phases) and pattern of enhancement. Enhancement patterns were characterized into 3 types: hyperenhanced, hypoenhanced, and unenhanced. The imaging results were compared with the histologic results. Contrast-enhanced time-intensity curves were used to assess the perfusion characteristics of the lesions by offline measurements with the Analysis QontraXt version 4.0 software. Regions of interest (ROIs) of similar size were drawn by each observer on the lesions and the normal circumrectal tissues carefully to avoid the lumen and areas of necrosis (appearing as contrast-filling defects). The timeintensity curve was obtained from the stored video clips after the ROIs were set. The ROI position was adjusted manually from frame to frame as necessary. The contrast arrival time, time to peak, peak intensity, sharpness, and area under the curve (AUC) within the rectal lesions and
Figure 1. Transrectal sonograms of a normal rectal wall visualized as a 5-layer echo structure (arrows). A, Before intraluminal contrast agent administration. B, After instillation of the contrast agent. The rectal lumen was dilated by the contrast agent, which can serve as a homogeneous acoustic window and can better delineate the wall of the rectum. A
J Ultrasound Med 2014; 33:575–583
B
577
3304jum557-712online_Layout 1 3/19/14 11:30 AM Page 578
Lu et al—Double–Contrast-Enhanced Sonography for Diagnosis of Rectal Lesions
the normal circumrectal tissues were computed by the built-in software. The parameter arrival time (in seconds) was defined as the first point of the curve clearly above the baseline intensity, followed by a rise. The time to peak (in seconds) was defined as the time from the start of the injection to the maximum intensity of the curve. The peak intensity (in decibels) was defined as the maximum intensity. The AUC can be directly estimated from the integral of the signal intensity versus time curve. Parameters obtained from 3 ROIs were averaged, and the values were taken as the representative parameters for a single time. A duplicate measurement was done by the same reviewer. Finally, the representative parameter values were the averages of the measurements by the 2 reviewers. Statistical Analyses Statistical analyses were performed with a standard statistical software package (SPSS version 18.0; IBM Corporation, Armonk, NY). Time-intensity curve perfusion parameters, including the arrival time, peak intensity, and AUC, were normally distributed continuous data and were expressed as mean ± standard deviation. Differences in time-intensity curve perfusion parameters between different lesions and normal perirectal tissue were compared by an independent samples t test. Comparisons of time-intensity curve parameters among the 3 rectal lesions and normal perirectal tissue were performed with 1-way analysis of variance, and a Student-Newman-Keuls test was conducted for pair-wise comparison of perfusion parameters among the different lesions and normal perirectal tissue. P < .05 was considered statistically significant.
Results Of 420 patients referred for double–contrast-enhanced sonography, 227 patients with 232 lesions (5 patients had 2 lesions) met the inclusion/exclusion criteria and were included in this study. There were 143 male and 84 female patients with a mean age of 60 ± 13 years. All 227 patients tolerated the examinations well with no immediate adverse events. The mean lesion size was 5.4 cm (range, 2.5–10 cm). Histopathologic examination showed 172 adenocarcinomas, 45 adenomas, and 15 inflammatory masses. Histopathologic stages of adenocarcinomas were 34 in T1, 56 in T2, 70 in T3, and 12 in T4. Eighty-one lesions were located in the lower third (0– 5 cm) of the rectum (7 inflammatory masses, 18 adenomas, and 56 adenocarcinomas), 86 in the middle third (5–8 cm; 4 inflammatory masses, 15 adenomas, and 67 adenocarcinomas), and 65 in the higher third (8–12 cm; 4
578
inflammatory masses, 12 adenomas, and 49 adenocarcinomas). There were 56 lymph node metastases among 114 total lymph nodes and 8 distant metastases. Double–contrast-enhanced sonography of adenocarcinomas showed a hyperenhanced homogeneous appearance in 43 of 172 lesions (25%; Figure 2), a hyperenhanced heterogeneous pattern in 112 of 172 (65%), and hypoenhanced and unenhanced patterns in 17 of 172 (10%). Seventeen of 45 adenomas showed a hypoenhanced homogeneous pattern, whereas 28 of 45 had a hyperenhanced homogeneous appearance on double– contrast-enhanced sonography (Figure 3). All 15 inflammatory lesions showed a hyperenhanced rim pattern without central enhancement (Figure 4). Table 1 shows comparisons between time-intensity curve perfusion parameters for different lesions and normal rectal tissue. Analysis of variance showed significant differences between different lesions (F = 87.72; P < .001). Adenocarcinomas had a faster arrival time and time to peak than normal rectal tissue, and the differences were significant (P < .05; Figure 2). Adenomas had a later time to peak and lower peak intensity than normal rectal tissue, and the differences were significant (P < .01; Figure 3). Inflammatory lesions had a faster arrival time and higher peak intensity than normal rectal tissue, and the differences were significant (P < .05; Figure 4). Statistically significant differences in perfusion parameters were seen between adenocarcinomas, adenomas, and inflammatory lesions (Table 1). Adenocarcinomas had a higher peak intensity and faster arrival time than adenomas, and the differences were significant (P < 0.001; Figure 5). Adenocarcinomas had a higher peak intensity and faster time to peak than inflammatory lesions, and the differences were significant (P < .01). Inflammatory masses had a faster arrival time, higher peak intensity, and later time to peak than adenomas, and the differences were significant (P < .05).
Discussion Colorectal cancer is the third most common cancer worldwide.1 Approximately 40% to 50% of colorectal cancers are located in the rectum. More than 90% of cases of rectal cancer develop from benign adenomas, but only approximately 10% of rectal adenomas will eventually develop into rectal cancer.7 Benign adenomas only require local excision with either endoscopic or surgical polypectomy techniques. For rectal lesions that are not amenable to endoscopic removal or in cases of uncertainty, accurate preoperative assessment of the lesion is important for coun-
J Ultrasound Med 2014; 33:575–583
3304jum557-712online_Layout 1 3/19/14 11:30 AM Page 579
Lu et al—Double–Contrast-Enhanced Sonography for Diagnosis of Rectal Lesions
seling the patient regarding the best options for removal, perioperative risks, and chance of recurrence.14,15 For a small but crucial group of patients, differentiation of a benign rectal mass from malignant neoplasia remains an unsolved problem despite improved technology. Furthermore, inflammatory masses are challenging to diagnose and differentiate from carcinomas. Standard sonography often is of little benefit, as inflammatory masses appear isoechoic or hypoechoic to the surrounding tissue, similar to rectal cancer. The difference is more easily distinguishable when the mass includes an abscess fluid pocket. To our knowledge, this study was the first to use a quantitative analysis of double–contrast-enhanced sonography for discrimination of rectal adenocarcinomas from
masslike adenomas and inflammatory masses. Chung et al16 demonstrated that a luminal contrast agent was a useful addition to sonographic examination of the bowel. It provided reliable evaluation of the bowel wall structure and improved the detection of wall changes compared to conventional sonography. In our study, we used a new luminal contrast agent to provide a homogeneous acoustic window. The luminal contrast agent provided excellent delineation of the 5-layer echo pattern of the normal rectal wall and the morphologic characteristics of rectal lesions. Villous rectal lesions appear as a polypoid or fat during endoscopy or standard transrectal sonography. In this study, 4 villous adenomas were observed and appeared to “float” within the contrast agent–filled lumen.
Figure 2. Images from a 63-year-old male patient with an adenocarcinoma in the middle third of the rectum. A, Endoscopic image showing the adenocarcinoma. B, Microscopic view of a hematoxylin-eosin–stained specimen showing the adenocarcinoma (original magnification ×100). C, After rectal instillation of the contrast agent, transrectal sonography with luminal extension showed a heterogeneous hypoechoic mass with irregular margins. D, After intravenous contrast agent injection, double–contrast-enhanced sonography in the harmonic imaging mode showed a hyperenhanced homogeneous adenocarcinoma with a clear margin and a time-intensity curve measurement indicating a low-perfusion pattern in the normal rectal wall. E, Regions of interest of the time-intensity curve were drawn within the lesion and a showed a high-perfusion pattern in the lesion.
J Ultrasound Med 2014; 33:575–583
579
3304jum557-712online_Layout 1 3/19/14 11:30 AM Page 580
Lu et al—Double–Contrast-Enhanced Sonography for Diagnosis of Rectal Lesions
Adenocarcinoma sonographically appears as an irregular and “fixed” or “immobile” lesion. In this study, double– contrast-enhanced sonography clearly demonstrated that even a very small adenocarcinoma (1 cm) appeared as a polypoid and fixed lesion within the contrast agent–filled cavity. Therefore, a gastrointestinal luminal contrast agent could improve the morphologic assessment of rectal lesions to help differentiate malignant from benign lesions. This study demonstrates that aggressive malignant tumors have a high degree of angiogenesis. In our experience, after intravenous administration of the contrast agent, adenocarcinomas displayed marked and heterogeneous signal enhancement in 90% of cases. However, as tumors increase in size, they may outgrow their blood supply, resulting in necrosis within the tumors.17,18 Larger tumors in our study tended to be hypoenhanced or unenhanced, indicating that the lesions had necrotic changes. These features were absent in inflammatory lesions. The enhancement pattern of the inflammatory lesions included a hyperenhanced rim appearance and no central enhancement. In this study, all adenomas had homogeneous enhancement, and most of the large adenomas had a hyperenhanced pattern. This preliminary study shows that
differences in enhancement patterns may help in the differentiation of rectal lesions. Double–contrast-enhanced sonography not only showed high-resolution imaging of the morphologic characteristics of rectal tumors but also provided useful, albeit more limited, quantitative parameters of tumor perfusion characteristics. Time-intensity curve parameters provide a quantitative means to differentiate benign from malignant rectal lesions. These data are easy to interpret and can be readily implemented into routine clinical practice. In this study, when compared to normal rectal tissue, we found that rectal adenocarcinomas had a faster arrival time and time to peak, adenomas had a lower peak intensity, and inflammatory lesions had an earlier arrival time and a higher peak intensity. Our study also showed that the vascular supply of rectal lesions differed significantly between benign and malignant lesions, which could be used as an adjunctive diagnostic tool in challenging cases. There were some limitations to this study. First, we did not study the relationship between the enhanced intensity revealed by double–contrast-enhanced sonography and microvessel density in rectal lesions. Further investi-
Figure 3. Images from a 47-year-old female patient with a rectal adenoma. A, Endoscopic image showing a tubulovillous adenoma with no malignant features. B, Microscopic view of a hematoxylin-eosin–stained specimen showing the tubulovillous adenoma (original magnification ×100). C, Transrectal sonography with luminal extension by the contrast agent showed a heterogeneous hypoechoic mass with regular margins in the lower third of the rectum. D, With intravenous contrast agent injection, double–contrast-enhanced sonography in the harmonic imaging mode showed a hyperenhanced homogeneous adenoma with a clear margin and a time-intensity curve measurement indicating a low-perfusion pattern in the lesion. TP indicates time to peak.
580
J Ultrasound Med 2014; 33:575–583
3304jum557-712online_Layout 1 3/19/14 11:30 AM Page 581
Lu et al—Double–Contrast-Enhanced Sonography for Diagnosis of Rectal Lesions
gation on this parameter may improve the diagnostic power of double–contrast-enhanced sonography. Second, the placement and depth of the ROI may have influenced the parameter results. We used the average of multiple different measurements from independent reviewers in an attempt to minimize the effect of ROI selections. Third, there is certain subjectivity to assessing morphologic enhancement, such as determining homogeneous versus heterogeneous patterns. Given that this modality is heavily operator dependant, accuracy may be reduced in nonexpert hands.
In conclusion, we have provided a preliminary report of evaluation and diagnosis of rectal lesions by using a new double–contrast-enhanced sonographic technique. Double–contrast-enhanced sonography can improve and overcome some limitations of traditional transrectal sonography by providing an optimal acoustic window for both grayscale and contrast imaging of the rectal lesions. Morphologic enhancement characteristics and quantitative perfusion parameters may help differentiate rectal adenocarcinomas from adenomas or inflammatory lesions. Thus, double–contrast-enhanced sonography may become a useful method for characterizing and evaluating the blood supply of rectal lesions to differentiate benign from malignant lesions.
Figure 4. Images from a 27-year-old male patient with an inflammatory mass in the lower third of the rectum. A, Endoscopic image showing the inflammatory lesion. B, Microscopic view of a hematoxylin-eosin–stained specimen showing the inflammatory lesion (original magnification ×100). C, Transrectal sonography with color Doppler showed a heterogeneous hypoechoic mass (calipers) with irregular margins in the lower third of the rectum. D, Double–contrast-enhanced sonography with intravenous contrast agent injection showed a hypoenhanced heterogeneous mass with a time-intensity curve measurement indicating very low perfusion inside the mass. E, The time-intensity curve measurement of the enhanced rim of the lesion showed relatively high perfusion compared to that of inside the lesion. TP indicates time to peak.
J Ultrasound Med 2014; 33:575–583
581
3304jum557-712online_Layout 1 3/19/14 11:30 AM Page 582
Lu et al—Double–Contrast-Enhanced Sonography for Diagnosis of Rectal Lesions
Table 1. Time-Intensity Curve Values for the Different Rectal Lesions Lesion
n
Adenocarcinoma 172 Adenoma 45 Inflammatory mass 15 Normal perirectal tissue 232 P (t test) Adenocarcinoma vs adenoma Adenocarcinoma vs inflammatory mass Adenocarcinoma vs normal perirectal tissue Adenoma vs inflammatory mass Adenoma vs normal perirectal tissue Inflammatory mass vs normal perirectal tissue
AT, s
TP, s
PI, dB
AUC
19.9 ± 3.5 24.4 ± 1.7 20.7 ± 3.4 24.1 ± 2.5
25.5 ± 77 25.2 ± 12.5 42.9 ± 26.1 41.4 ± 17.2
46.4 ± 9.2 16.5 ± 5.9 62.0 ± 8.9 44.3 ± 8.9
6.6 ± 3.9 4.0 ± 2.9 3.8 ± 4.8 4.2 ± 3.3
.99 .99 .013
>.99 .007 .035 .008 .035 >.99
