DOI: 10.1148/rg.283075188
RadioGraphics 2008;28:623-638
© RSNA, 2008
Quality Initiatives
Missed Lesions at Abdominal Oncologic CT: Lessons Learned from Quality Assurance1
Bettina Siewert, MD,
Jacob Sosna, MD,
Ann McNamara, MD,
Vassilios Raptopoulos, MD, and
Jonathan B. Kruskal, MD, PhD
1 From the Department of Radiology, Beth Israel Deaconess Medical Center, 330 Brookline Ave, Boston, MA 02215. Presented as an education exhibit at the 2003 RSNA Annual Meeting. Received September 12, 2007; revision requested December 19 and received February 5, 2008; accepted February 21. J.S. receives research support from Koninklijke Philips Electronics; V.R. receives research support from Toshiba and E-Z-Em; all other authors have no financial relationships to disclose.
Address correspondence to B.S. (e-mail: bsiewert{at}bidmc.harvard.edu).
 |
Abstract
|
|---|
The evaluation of oncology patients represents a substantial volume of the workload in many radiology departments. Interpreting the results of oncologic examinations is often challenging and time-consuming because many abnormalities are identified in the same examination and must be compared with the findings in previous studies. However, errors in the interpretation of oncologic computed tomographic (CT) scans can have significant effects on patient care. These effects may range from withdrawal from a clinical trial or cessation of therapy to repeat CT examination because of a technically inadequate study, CT-guided biopsy of newly identified lesions, or initiation of therapy for previously unrecognized lesions. A root cause analysis of reported errors in the interpretation of abdominal and pelvic CT scans led to the identification of potential pitfalls that may be encountered when interpreting oncologic CT scans and factors that contribute to these errors. Awareness of the spectrum of factors that contribute to misinterpretation of CT scans in oncology patients may improve the performance of the individual radiologist and ultimately translate into improved patient care.
© RSNA, 2008
|
Introduction
|
|---|
Imaging evaluation of patients with cancer makes up a substantial volume of the workload in most hospital radiology departments (1). This load may be further compounded by trials evaluating the efficacy of experimental chemotherapies that require multiple tumor measurements and comparisons to be made with prior studies for assessment of therapeutic response. Meeting the stringent inclusion criteria and measurement requirements of national oncology groups in patients with multivisceral lesions adds to the complexity and time required to interpret results of these studies.
Discordant interpretations of computed tomographic (CT) scans are common and have been reported in 31%–37% of cases (2–4). Major discrepancies have been reported in up to 17% of cases (2), resulting in a change in radiologic staging in 19% (4). One study showed that management was affected in 7% of cases with a change in treatment being initiated in 4% of patients (4). This was confirmed by Gollub et al (2), who reported an actual change of treatment in 3%. However, more recent multicenter trials have reported changes in patient treatment in as many as 23% of patients due to discordant readings (5). Clearly, discordant readings affect patient care and treatment.
A change in patient treatment may affect outcome, whether this is positive or adverse. This is especially relevant to patients being evaluated for follow-up of malignancies. Some of the changes in patient care may include withdrawal from a clinical trial, cessation of therapy, repeat CT examination for a technically inadequate study, CT-guided biopsy for newly identified lesions, or initiation of therapy for previously unrecognized lesions. The emotional impact on the patient and the associated frustration and additional work this may cause the treating physician must also be considered.
Bechtold et al (3) investigated errors in interpretation of abdominal CT scans and their causes. Surprisingly, a number of factors did not influence the error rate, including simultaneous supervision of interventional procedures (3), level of training of assigned resident physicians or fellows (3), or tumor type (4). Although the error rate doubled from 7% (if scans from < 20 studies were read per day) to 15% (if scans from > 20 studies were read per day) (3), this difference was not statistically significant. 
The only factor in this study to reach statistical significance was the skill of the individual radiologist, with error rates ranging from 3.6% to 16.1% (mean error rate, 7.6%).
On the basis of our root cause analysis of over 250 reported errors in the interpretation of abdominal and pelvic CT scans, the intention of this review is to make the reader aware of potential pitfalls that may be encountered when interpreting oncologic CT scans and factors that contribute to these errors. Awareness of the spectrum of factors that contribute to misinterpretation of CT scans in oncology patients may improve the performance of the individual radiologist and ultimately translate into improved patient treatment.
|
Categories of Errors on Oncologic CT Scans
|
|---|
What is an error? Errors may be categorized in a variety of ways. We have methods in place to facilitate identification of errors so that steps can be introduced to minimize their occurrence. In addition, errors can be further classified in terms of the outcome or harm suffered by the patient, if any, and for assessing and assigning accountability of the person involved.
In broad terms, factors contributing to errors are categorized as being system related (latent errors) or person related (active errors) (6). These human cognitive errors are more likely to be preventable and more likely to have an adverse outcome than technical errors. As applied to diagnostic radiology, three main categories of error are responsible for the majority of "missed" or misinterpreted observations on oncologic CT scans: technical (latent or system related), active (errors in perception, in knowledge, in judgment), or a combination thereof.
|
Technical Errors
|
|---|
Scanning Factors
In a study by Loughrey et al (4) that reviewed oncologic CT scans, 94% of studies were deemed technically adequate. Standardized protocols depending on tumor type should be employed to allow optimal visualization of metastatic disease in organs that are likely to be affected by the primary tumor. While studies are usually performed with oral and intravenous contrast material (including preliminary nonenhanced images in many cases), image acquisition after contrast material injection varies from arterially enhanced scans only to portal venous phase and delayed imaging or may include all series as part of a multiphase examination. It is very important that studies are performed with full knowledge of the primary tumor. For example, unless delayed images are obtained, insufficient diagnostic or staging information can be provided about cholangiocarcinoma (7). Conversely, neuroendocrine tumors tend to have early arterial enhancement and therefore comparison should be made between the arterial phase of each study. Unless images are obtained during the bolus phase of contrast material administration, vascular lesions or metastases may easily be missed, resulting in understaging of a malignancy.
Scanning Parameters
Imaging Protocols.—
The selected protocols vary depending on tumor type and may include nonenhanced, bolus phase, non–equilibrium phase, and delayed imaging; all of these are timed according to the primary neoplasm and anatomic regions of interest. The Table provides a summary of oncology protocols used at our institution. Oncology protocols most commonly use a section thickness of 5 mm when follow-up imaging is performed. For staging protocols that require analysis of vascular invasion, thinner sections (down to 0.625 mm and depending on available CT scan technology) are required. Two-dimensional and three-dimensional reconstruction can be valuable and time efficient. In particular, when surveying the osseous structures for metastatic disease, our experience is that images reformatted in the sagittal and coronal planes are essential for evaluating bones for possible metastatic disease, and they have become part of our routine protocols.
Use of Intravenous Contrast Material.—
Intravenous contrast material is routinely used, with the amount depending on the patient’s weight (100 mL for patients < 155 lb [< 70 kg], 130 mL for patients > 155 lb [> 70 kg]). The injection rate for routine protocols is 2.5 mL/sec. However, certain groups of patients are not likely to be administered intravenous contrast material. Patients with impaired renal function who have already undergone a nephrectomy are not usually given intravenous contrast material. Subtle metastases from renal cell carcinoma may thus be missed in areas such as the pancreas, where the early enhancement of renal cell carcinoma metastases typically facilitates lesion identification.
Body Habitus.—
Oncologic imaging has to be adjusted to the patient’s body habitus. In obese patients, scan parameters have to be modified to provide sufficient signal-to-noise ratio. Our routine protocols are performed with 120 kVp and 320 mAs. In general, an increase in kilovolt peak to 140 will sufficiently increase the effective milliampereseconds value and provide less image noise (8). However, these adjustments will increase the patient’s radiation dose. In addition, attention has to be paid to the subcutaneous soft tissues in these patients. While the temptation exists to perform aggressive image cropping in an attempt to focus on the internal organs, important information in the soft tissues (such as metastatic deposits) will get lost; this may be especially important in patients with metastatic melanoma.
Dose Reduction.—
Recently, the use of low-dose protocols has gained considerable attention. Substantial dose reduction up to 64% (9) without a loss in information has been shown in patients with renal colic (9) or urolithiasis (10), those undergoing CT angiography for living related kidney donation (11), and for quantification of emphysema (12). Oncologic patients undergoing routine surveillance CT examinations several times per year are among those who would benefit most from dose reduction strategies. However, to our knowledge there are currently no published studies comparing routine oncologic CT to low-dose techniques.
Soft-Tissue Opacification
In general, CT protocols for evaluation of oncology patients should be performed with intravenous contrast material, increasing the contrast differential between normal tissue and tumor. It thus facilitates the depiction of more tumor foci (Fig 1). Because the borders of a mass are better outlined on a contrast-enhanced scan, more accurate tumor measurements can be performed and evaluation for tumor response to treatment is improved. Our routine protocols use an automated triggering approach with region-of-interest placement over the hepatic parenchyma with an automatic trigger once the liver parenchyma reaches 50 HU (13). Care has to be taken when positioning the region of interest to avoid inadvertently placing this over a vessel or a focal liver lesion, which would alter study quality. However, optimal contrast between normal liver tissue and metastatic disease depends on the histologic features of the tumor. Nonenhanced images remain of value for detection of lesions that demonstrate early enhancement with intravenous contrast material and may thus be indistinguishable from normal parenchyma on portal venous images (14), in particular in patients with breast cancer if no arterial phase imaging is performed.
View larger version (117K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1a. Tumor recurrence in an 83-year-old man with prostate cancer. (a) On a CT scan, it is difficult to detect a small recurrent tumor (arrow) in the prostate bed. The patient presented 11 months later with hematuria. (b) On a nonenhanced image, it is difficult to distinguish the tumor (arrow) from the pelvic floor muscles. (c) On a contrast-enhanced image, the lesion (arrow) is easily identified. (d) Coronal reformatted image shows infiltration of the bladder (arrow).
|
|
View larger version (121K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1b. Tumor recurrence in an 83-year-old man with prostate cancer. (a) On a CT scan, it is difficult to detect a small recurrent tumor (arrow) in the prostate bed. The patient presented 11 months later with hematuria. (b) On a nonenhanced image, it is difficult to distinguish the tumor (arrow) from the pelvic floor muscles. (c) On a contrast-enhanced image, the lesion (arrow) is easily identified. (d) Coronal reformatted image shows infiltration of the bladder (arrow).
|
|
View larger version (122K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1c. Tumor recurrence in an 83-year-old man with prostate cancer. (a) On a CT scan, it is difficult to detect a small recurrent tumor (arrow) in the prostate bed. The patient presented 11 months later with hematuria. (b) On a nonenhanced image, it is difficult to distinguish the tumor (arrow) from the pelvic floor muscles. (c) On a contrast-enhanced image, the lesion (arrow) is easily identified. (d) Coronal reformatted image shows infiltration of the bladder (arrow).
|
|
View larger version (112K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1d. Tumor recurrence in an 83-year-old man with prostate cancer. (a) On a CT scan, it is difficult to detect a small recurrent tumor (arrow) in the prostate bed. The patient presented 11 months later with hematuria. (b) On a nonenhanced image, it is difficult to distinguish the tumor (arrow) from the pelvic floor muscles. (c) On a contrast-enhanced image, the lesion (arrow) is easily identified. (d) Coronal reformatted image shows infiltration of the bladder (arrow).
|
|
Vascular Opacification
Intravenous contrast material is invaluable in differentiating lymphadenopathy from poorly opacified or nonenhanced vessels (Fig 2). This is of particular importance in the pelvis, where a large number of vessels are routinely encountered in a small anatomic region where nodes may coexist. Images in the venous phase are thus particularly helpful because the increased attenuation in the vessel allows easier distinction from lymph nodes (Figs 3, 4). A combination of careful visualization, familiarity with expected pathways of nodal tumor spread, and reviewing images in coronal reformatted planes may enhance the likelihood of detecting enlarged nodes.
View larger version (116K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2a. Lymphadenopathy in a 63-year-old woman with a history of lymphoma. (a) On a CT scan, it is difficult to differentiate left iliac lymphadenopathy (arrow) from the external iliac vein owing to their similar attenuation. (b) On an image obtained at 9-month follow-up, the lesion (arrow) is more conspicuous due to an increase in size and slightly different attenuation from that of the vein.
|
|
View larger version (105K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2b. Lymphadenopathy in a 63-year-old woman with a history of lymphoma. (a) On a CT scan, it is difficult to differentiate left iliac lymphadenopathy (arrow) from the external iliac vein owing to their similar attenuation. (b) On an image obtained at 9-month follow-up, the lesion (arrow) is more conspicuous due to an increase in size and slightly different attenuation from that of the vein.
|
|
View larger version (121K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3a. Lymphadenopathy in a 34-year-old woman with renal cell carcinoma. (a) On a CT scan, it is difficult to differentiate retrocaval lymphadenopathy (arrow) from the inferior vena cava. (b) On a delayed image, the inferior vena cava (arrowhead) is well opacified and the lymphadenopathy (arrow) is more conspicuous.
|
|
View larger version (119K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3b. Lymphadenopathy in a 34-year-old woman with renal cell carcinoma. (a) On a CT scan, it is difficult to differentiate retrocaval lymphadenopathy (arrow) from the inferior vena cava. (b) On a delayed image, the inferior vena cava (arrowhead) is well opacified and the lymphadenopathy (arrow) is more conspicuous.
|
|
View larger version (113K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4a. Lymph node metastasis in a 73-year-old man with prostate cancer. (a) On a CT scan, a metastasis in a left iliac lymph node (arrow) is difficult to differentiate from a pelvic vein owing to their similar enhancement. (b) On a 6-month follow-up image obtained with slightly different timing of the contrast material injection, the node (arrow) is more conspicuous in comparison with the now opacified vein (arrowhead).
|
|
View larger version (111K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4b. Lymph node metastasis in a 73-year-old man with prostate cancer. (a) On a CT scan, a metastasis in a left iliac lymph node (arrow) is difficult to differentiate from a pelvic vein owing to their similar enhancement. (b) On a 6-month follow-up image obtained with slightly different timing of the contrast material injection, the node (arrow) is more conspicuous in comparison with the now opacified vein (arrowhead).
|
|
Bladder Opacification
In patients with a history of a malignancy of the genitourinary tract (in particular bladder and prostate carcinoma), opacification of the bladder is important and delayed images of the pelvis must be obtained no sooner than 10 minutes after the start of intravenous contrast material injection to allow adequate distention and opacification of the bladder (Fig 5). Interpretation of the surgical field is generally more challenging because postsurgical changes lead to increased attenuation of soft tissues as well as anatomic distortion, which may arise from scarring, radiation therapy, surgical clips, and even orthopedic hardware. Differentiation of early local tumor recurrence from postsurgical scarring is challenging and can often be achieved only in subsequent follow-up examinations. However, use of intravenous contrast material to enhance attenuation differences between soft tissues and tumor (Fig 1) as well as complete bladder opacification can improve the sensitivity of CT.
View larger version (111K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5a. Soft-tissue mass in a 58-year-old man with bladder cancer who underwent cystectomy and creation of a neobladder. (a) On a CT scan, it is difficult to detect a small soft-tissue mass (arrow) at the anastomosis of the neobladder with the urethra. (b) On a follow-up image obtained 10 months later, it is still difficult to detect the mass (arrow) owing to lack of opacification of the neobladder (arrowhead). (c) On a delayed image obtained after satisfactory opacification of the bladder (arrowhead), the mass (arrow) is easily detected.
|
|
View larger version (112K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5b. Soft-tissue mass in a 58-year-old man with bladder cancer who underwent cystectomy and creation of a neobladder. (a) On a CT scan, it is difficult to detect a small soft-tissue mass (arrow) at the anastomosis of the neobladder with the urethra. (b) On a follow-up image obtained 10 months later, it is still difficult to detect the mass (arrow) owing to lack of opacification of the neobladder (arrowhead). (c) On a delayed image obtained after satisfactory opacification of the bladder (arrowhead), the mass (arrow) is easily detected.
|
|
View larger version (108K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5c. Soft-tissue mass in a 58-year-old man with bladder cancer who underwent cystectomy and creation of a neobladder. (a) On a CT scan, it is difficult to detect a small soft-tissue mass (arrow) at the anastomosis of the neobladder with the urethra. (b) On a follow-up image obtained 10 months later, it is still difficult to detect the mass (arrow) owing to lack of opacification of the neobladder (arrowhead). (c) On a delayed image obtained after satisfactory opacification of the bladder (arrowhead), the mass (arrow) is easily detected.
|
|
Bowel Opacification
For certain malignancies, especially those likely to metastasize to the bowel, complete opacification of the gastrointestinal tract is desirable because metastases to the small bowel (Fig 6) and large bowel (Figs 7, 8) are difficult to detect, especially if the bowel is not distended (Fig 7). In addition, mesenteric lymphadenopathy (Fig 9) and omental masses can be mistaken for normal-sized unopacified bowel (4). Our routine protocols, performed after patient ingestion of two bottles of a 20% barium solution (Barocat; E-Z-Em, New York, NY), provide satisfactory bowel opacification; however, complete opacification can be obtained with three bottles. Alternatively, neutral and negative oral contrast media can be used (15,16). Gastrointestinal stromal tumors as well as lymphoma and melanoma may produce cavitary masses, and good bowel opacification may aid in their detection. In the absence of oral contrast material, metastases that demonstrate central necrosis are particularly easily mistaken for normal bowel, as the outer enhancing rim can simulate normal enhancement of the bowel wall and the area of central necrosis may be thought to represent fluid within the bowel lumen.
View larger version (130K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6a. Small bowel metastasis in a 53-year-old woman with colon cancer. (a) On a CT scan, it is difficult to perceive a metastasis to the small bowel (arrow) owing to lack of oral contrast material and abnormal adjacent bowel loops with wall thickening and mural edema (arrowhead), which are likely due to ischemia. (b) On an image obtained at 2-month follow-up, the mass (arrow) is easily distinguished from the loops of small bowel (arrowhead), which are now well opacified with oral contrast material.
|
|
View larger version (123K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6b. Small bowel metastasis in a 53-year-old woman with colon cancer. (a) On a CT scan, it is difficult to perceive a metastasis to the small bowel (arrow) owing to lack of oral contrast material and abnormal adjacent bowel loops with wall thickening and mural edema (arrowhead), which are likely due to ischemia. (b) On an image obtained at 2-month follow-up, the mass (arrow) is easily distinguished from the loops of small bowel (arrowhead), which are now well opacified with oral contrast material.
|
|
View larger version (117K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 7a. Large bowel mass in an 80-year-old woman with colon cancer. (a) On a CT scan, it is difficult to discern a mass in the ascending colon (arrow) owing to lack of oral contrast material. (b) On an image obtained at 2-month follow-up, the lesion (arrow) is easily demonstrated due to adequate bowel opacification.
|
|
View larger version (125K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 7b. Large bowel mass in an 80-year-old woman with colon cancer. (a) On a CT scan, it is difficult to discern a mass in the ascending colon (arrow) owing to lack of oral contrast material. (b) On an image obtained at 2-month follow-up, the lesion (arrow) is easily demonstrated due to adequate bowel opacification.
|
|
View larger version (124K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 8a. Large bowel metastasis in an 87-year-old woman with a history of colon cancer. (a) On a CT scan, it is difficult to detect a metastasis to the descending colon (arrow) owing to lack of oral contrast material (arrowhead). (b) On an image obtained at 6-month follow-up, the lesion is slightly increased in size (arrow) but is difficult to detect due to lack of oral and intravenous contrast material. The lack of contrast material makes it difficult to distinguish the lesion from a fluid-filled bowel loop.
|
|
View larger version (135K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 8b. Large bowel metastasis in an 87-year-old woman with a history of colon cancer. (a) On a CT scan, it is difficult to detect a metastasis to the descending colon (arrow) owing to lack of oral contrast material (arrowhead). (b) On an image obtained at 6-month follow-up, the lesion is slightly increased in size (arrow) but is difficult to detect due to lack of oral and intravenous contrast material. The lack of contrast material makes it difficult to distinguish the lesion from a fluid-filled bowel loop.
|
|
View larger version (131K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 9a. Mesenteric lymphadenopathy in a 50-year-old man with a history of renal cell cancer. (a) On a CT scan, mesenteric lymphadenopathy (arrow) is not recognizable due to suboptimal bowel opacification (arrowhead). (b) On an image obtained at 2-month follow-up, the lymphadenopathy is easier to recognize (arrow) due to the complete bowel opacification (arrowhead).
|
|
View larger version (127K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 9b. Mesenteric lymphadenopathy in a 50-year-old man with a history of renal cell cancer. (a) On a CT scan, mesenteric lymphadenopathy (arrow) is not recognizable due to suboptimal bowel opacification (arrowhead). (b) On an image obtained at 2-month follow-up, the lymphadenopathy is easier to recognize (arrow) due to the complete bowel opacification (arrowhead).
|
|
Other technical factors that may contribute to lesions being missed include patient motion, peristalsis of bowel, and the inability to breath hold for prolonged periods of time. Also, volume averaging and section reconstruction algorithms should be considered.
|
Active Errors
|
|---|
Perceptual Factors
Systematic viewing of all organs in a standardized sequence ensures that areas are not overlooked (17). This includes the use of several different window settings and awareness of "blind spots," which should be analyzed carefully. In addition, the reader has to be aware of premature satisfaction of search, a common problem in oncology examination where often multiple observations have to be made in a single study and comparison of these with prior imaging results may be tedious. With the increase in study volume, it is of utmost importance that reading stations are ergonomically sound to avoid early reader fatigue.
Window Settings.—
The use of additional window settings, particularly in evaluating the liver (Figs 10, 11) and bones (Fig 12), is very helpful (1). Regular soft-tissue windows apply a window width of 350 HU and window level of 50 HU. Liver windows with a window width of 150 HU and window level of 100 HU improve soft-tissue contrast between normal liver tissue and metastases, making them more conspicuous. Dedicated bone windows (window width, 3000 HU; window level, 500 HU) are critical for the detection of osseous metastases. We suggest the following algorithm for an oncologic CT study of the torso: chest images with a soft-tissue window, chest images with a lung window (window width, 1500 HU; window level, -500 HU), abdominal images (all phases of enhancement) with a soft-tissue window, abdominal images (all phases of enhancement) with a liver window, pelvic images with a soft-tissue window, and images of the entire torso with a bone window.
View larger version (153K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 10a. Liver metastases in a 76-year-old woman with renal cell carcinoma. (a) On a nonenhanced image displayed with a soft-tissue window, two liver metastases (arrows) are barely perceptible. (b) On an image displayed with a liver window, the lesions (arrows) are better visualized.
|
|
View larger version (100K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 10b. Liver metastases in a 76-year-old woman with renal cell carcinoma. (a) On a nonenhanced image displayed with a soft-tissue window, two liver metastases (arrows) are barely perceptible. (b) On an image displayed with a liver window, the lesions (arrows) are better visualized.
|
|
View larger version (140K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 11a. Liver metastases in a 66-year-old woman with endometrial carcinoma. (a) On an image displayed with a soft-tissue window, it is difficult to perceive a liver metastasis (arrow) in segment 2. (b) On an image displayed with a liver window, the lesion (arrow) is more conspicuous. (c) On an image obtained at 7-month follow-up, a second lesion (arrow) is visualized in segment 8.
|
|
View larger version (103K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 11b. Liver metastases in a 66-year-old woman with endometrial carcinoma. (a) On an image displayed with a soft-tissue window, it is difficult to perceive a liver metastasis (arrow) in segment 2. (b) On an image displayed with a liver window, the lesion (arrow) is more conspicuous. (c) On an image obtained at 7-month follow-up, a second lesion (arrow) is visualized in segment 8.
|
|
View larger version (88K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 11c. Liver metastases in a 66-year-old woman with endometrial carcinoma. (a) On an image displayed with a soft-tissue window, it is difficult to perceive a liver metastasis (arrow) in segment 2. (b) On an image displayed with a liver window, the lesion (arrow) is more conspicuous. (c) On an image obtained at 7-month follow-up, a second lesion (arrow) is visualized in segment 8.
|
|
View larger version (110K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 12a. Bone metastases in a 47-year-old woman with breast cancer. (a) On an image displayed with a soft-tissue window, it is impossible to detect bone metastases. (b) On an image displayed with a bone window, small sclerotic lesions (arrows) are seen in the lumbar spine, an appearance suggestive of metastatic disease. (c) On an image obtained at 3-month follow-up, the osseous lesions (arrows) are larger.
|
|
View larger version (86K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 12b. Bone metastases in a 47-year-old woman with breast cancer. (a) On an image displayed with a soft-tissue window, it is impossible to detect bone metastases. (b) On an image displayed with a bone window, small sclerotic lesions (arrows) are seen in the lumbar spine, an appearance suggestive of metastatic disease. (c) On an image obtained at 3-month follow-up, the osseous lesions (arrows) are larger.
|
|
View larger version (87K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 12c. Bone metastases in a 47-year-old woman with breast cancer. (a) On an image displayed with a soft-tissue window, it is impossible to detect bone metastases. (b) On an image displayed with a bone window, small sclerotic lesions (arrows) are seen in the lumbar spine, an appearance suggestive of metastatic disease. (c) On an image obtained at 3-month follow-up, the osseous lesions (arrows) are larger.
|
|
Blind Spots.—
Several problem areas have been described where lesions are most commonly overlooked. Awareness of these problem zones and dedicated analysis of these areas is recommended. These areas include subcutaneous and other soft tissues (Fig 13) such as the breast, the supraclavicular region and paraspinal region (Figs 14–16), and bones, in particular the scapula (3). The search for subcutaneous nodules should be particularly diligent in patients with a history of melanoma, where soft-tissue metastases are most common. Intramuscular metastases are problematic because metastases are often of similar attenuation to muscle and early diagnosis therefore relies on other subtle findings, such as contour abnormalities (Fig 16) and distortion of normal internal architecture, such as loss of internal septations and replacement of normal fatty tissue (Fig 14). Analysis of the supraclavicular region is challenging due to the complex anatomy in this region.
View larger version (105K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 13a. Obturator muscle metastasis in a 47-year-old man with renal cell carcinoma. (a) On a CT scan, it is difficult to discern a metastasis in the left obturator muscle (arrow). Note the subtle mass effect on the muscle with slight medial displacement. (b) Contrast-enhanced image shows some enhancement in the inferior aspect of the mass (arrow), which makes the mass more conspicuous.
|
|
View larger version (110K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 13b. Obturator muscle metastasis in a 47-year-old man with renal cell carcinoma. (a) On a CT scan, it is difficult to discern a metastasis in the left obturator muscle (arrow). Note the subtle mass effect on the muscle with slight medial displacement. (b) Contrast-enhanced image shows some enhancement in the inferior aspect of the mass (arrow), which makes the mass more conspicuous.
|
|
View larger version (109K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 14a. Paraspinal muscle metastasis in a 66-year-old man with a history of melanoma. (a) On a CT scan, a soft-tissue metastasis in the left paraspinal muscles (arrow) is inconspicuous due to its similar attenuation to that of muscle. Note the subtle distortion of the muscle architecture in comparison with that on the other side. (b) On an image obtained at 2-month follow-up, the lesion (arrow) is slightly larger and more conspicuous due to rim enhancement.
|
|
View larger version (115K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 14b. Paraspinal muscle metastasis in a 66-year-old man with a history of melanoma. (a) On a CT scan, a soft-tissue metastasis in the left paraspinal muscles (arrow) is inconspicuous due to its similar attenuation to that of muscle. Note the subtle distortion of the muscle architecture in comparison with that on the other side. (b) On an image obtained at 2-month follow-up, the lesion (arrow) is slightly larger and more conspicuous due to rim enhancement.
|
|
View larger version (110K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 15. Paraspinal mass in a 76-year-old woman with a history of lymphoma and renal cell carcinoma. The patient presented with new back pain. On a CT scan, it is difficult to detect a left paraspinal mass (arrow) owing to its similar attenuation to that of muscle. Again, note the subtle distortion of the muscle architecture.
|
|
View larger version (76K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 16a. Paraspinal metastasis in a 58-year-old man with renal cell carcinoma. (a) On a CT scan, it is difficult to discern a small left paraspinal metastasis (arrow). (b) On an image obtained at 2-month follow-up, the lesion (arrow) is easier to detect due to an area of central necrosis.
|
|
View larger version (80K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 16b. Paraspinal metastasis in a 58-year-old man with renal cell carcinoma. (a) On a CT scan, it is difficult to discern a small left paraspinal metastasis (arrow). (b) On an image obtained at 2-month follow-up, the lesion (arrow) is easier to detect due to an area of central necrosis.
|
|
With its combined anatomic and physiologic capabilities, fluorodeoxyglucose positron emission tomography (PET)/CT is commonly used for tumor staging. Experience gained with this hybrid imaging modality significantly enhances CT reading abilities: commonly overlooked regions and pathologic conditions are reviewed with more caution. A radiologist accustomed to reading PET/CT scans can become a better reader of multidetector CT scans, with lower probability of overlooking "easy to miss" pathologic conditions.
Picture Archiving and Communication Systems.—
Image review on picture archiving and communication system (PACS) workstations is very helpful, in particular in differentiating metastases from tubular structures such as vessels or bowel. Continuous scrolling through the area of abnormality allows one to follow a vessel over its course and to easily separate it from adjacent lymph nodes. Mesenteric lymphadenopathy or omental deposits can be better distinguished from bowel loops when the bowel is followed in a similar fashion.
However, certain issues with PACS systems have to be considered when oncologic scans are interpreted. These include the series directory, number of monitors, monitor setup, type of scroll function used, and status of the examination.
In contrast to hard copy images, where all series are printed on contiguous sheets, PACS usually display one series at a time and additional series can be accessed only via a directory. It is therefore important, at the end of each interpretation, to double-check the series directory for images that inadvertently may not have been viewed.
Two monitors are generally sufficient for review of oncologic study results. Four-monitor systems are most often used for interpretation of plain film images. However, on a four-monitor system, images from CT examinations may be displayed in a nonstandard way and the display on each monitor has to be double-checked prior to interpretation of the study results. Similarly, the display of images from prior and recent studies on a two-monitor display has to be double-checked to establish the correct time course of disease and avoid misinterpretation.
The type of scroll function can be selected as "diagnostic quality" or "image navigation." The image navigation display allows faster scrolling through an image set; however, it will cause blurring of the images and thus can make lesions less conspicuous. This can be avoided by using the diagnostic quality scroll function or paging through the images one at a time.
Finally, results of examinations should be interpreted only once they have been marked by the technologist as having been completed. Otherwise, series can be added to the examination on the PACS after image interpretation has been completed without the radiologist knowing that additional series were performed. This occurs most commonly when two-dimensional or three-dimensional reformations are performed with some time delay after the initial image acquisition. On rare occasions, depending on the PACS privileges of the user, image series may be added to an examination with a verified, dictated, or completed status without the radiologist being aware. Before interpreting the results of a study, it is therefore helpful to confirm the status of an examination as completed and to double-check under the series directory that all standard reformations have been performed.
Reading Environment and Ergonomics.—
With the advent of multidetector scanning, the reading environment has significantly changed. Study results are acquired faster, allowing us to obtain many image sequences over a short period of time, increasing the number of images from less than 100 for an examination of the abdomen and pelvis into the thousands. This represents an inordinate stress to the reader’s eyes and leads to faster reader fatigue. With the dramatic increase in the number of CT images that are now acquired, the interpreting radiologist is becoming exposed to a variety of ergonomic stresses that did not exist in the days of hard copy image review. As examples, mouse-induced carpal tunnel syndrome, eye strain, and even musculoskeletal injuries due to repetitive stress are being encountered with increasing frequency. Newly designed reading rooms are being constructed with an eye toward minimizing these stresses, with attention being focused on chair, workstation, and mouse engineering; correct posture during interpretation sessions; and minimizing ambient light and sound. A reduction in these stresses must certainly optimize the reading experience and contribute to limiting the possibility of errors.
Thus, ergonomics in the reading environment have become increasingly important. A variety of mouse types and chairs are available to meet the individual needs of the reader and ensure the most functional reading environment (18).
Satisfaction of Search.—
Satisfaction of search is defined as a phenomenon in which the detection of one abnormality interferes with the detection of other abnormalities (19). The interpretation of oncologic studies is particularly challenging because often multiple abnormalities coexist in a single study. It is of utmost importance to maintain a high level of suspicion for the entire duration of analysis and to not let attention slip after multiple lesions have already been diagnosed. In particular, when results of examinations with multiple preexisting lesions are compared, the temptation exists to not actively pursue the identification of new lesions.
Interpretative Factors
In order to avoid interpretative errors, the reader should be aware of the indications for the study, what therapies have been employed, and the spectrum of potential pitfalls that exist.
Study Indication.—
The indication for the study is important: requests for comparison studies may result in attention being directed to changes in preexisting lesions, resulting in new lesions or complications from lesions (such as pathologic fractures) being missed. In addition, the arduous task of measuring multiple lesions in oncologic follow-up studies may focus the reader’s attention away from three broad groups of additional findings that may require treatment: (a) complications of therapy (such as abscesses, fistulas, typhlitis, or other infections in immunocompromised patients); (b) complications of the disease process itself (such as vascular occlusion, hollow viscus perforation, bowel obstruction, or hemorrhage); and (c) other unexpected but common clinical findings, such as obstructing renal calculi or diverticulitis.
Effects of Therapies and Other Interventions.—
Certain therapies produce well-recognized anatomic changes in patients with malignancies. Perhaps the best example is the effect that combination chemotherapies have on the liver in patients being treated for breast cancer (20). Specifically, the fibrotic response that these drugs produce in the liver causes features of cirrhosis with regenerative nodules that should not be mistaken for hepatic metastases. With the broad clinical introduction of antiangiogenic therapies, the antitumor mechanisms of therapy are changing; one should expect to see a therapeutic response manifesting initially as a reduction in tumor vascularity rather than a decrease in tumor size. Thus, tumor stability may be wrongly assigned to solid masses that are in fact becoming hypo- or even avascular with treatment.
The imaging findings resulting from minimally invasive interventions may be misinterpreted. For example, readers should be well aware of expected perfusion and enhancement patterns that occur in solid organs after percutaneous radiofrequency ablation (21). In patients undergoing thermal ablative therapies, the larger coagulated zone should not be mistaken for an enlarging mass. Similarly, deposition of lipid vectors after intraarterial chemoembolization should not be misinterpreted as calcifications that may arise in mucinous colorectal cancer metastases to the liver. As more and more therapeutic interventions are introduced, including focused radiofrequency and microwave ablation, readers must be aware of the expected imaging findings in their patients.
Interpretive Pitfalls.—
Recognized pitfalls may result in findings being overcalled (false positive) or undercalled (false negative). The reader should also never forget that benign lesions may coexist with malignant tumors. Several pitfalls have been recognized for the interpretation of body CT scans (22,23). When contrast material is not used, aberrant vessels such as an aberrant subclavian artery (Fig 17) can be mistaken for mediastinal lymphadenopathy (4). This can be avoided by following the course of the structure throughout the mediastinum and establishing the connection with the aorta. Hypoattenuating lesions adjacent to the falciform ligament commonly represent focal fatty infiltration (Fig 18) and should not be mistaken for metastatic disease (4).
View larger version (84K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 17a. Aberrant subclavian artery mimicking mediastinal lymphadenopathy. (a) Nonenhanced CT scan shows a soft-tissue structure (arrow) adjacent to the esophagus. The soft-tissue structure could be mistaken for mediastinal lymphadenopathy. (b) Image obtained caudad to a shows a connection with the aorta (arrow), a finding that confirms the presence of an aberrant right subclavian artery.
|
|
View larger version (75K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 17b. Aberrant subclavian artery mimicking mediastinal lymphadenopathy. (a) Nonenhanced CT scan shows a soft-tissue structure (arrow) adjacent to the esophagus. The soft-tissue structure could be mistaken for mediastinal lymphadenopathy. (b) Image obtained caudad to a shows a connection with the aorta (arrow), a finding that confirms the presence of an aberrant right subclavian artery.
|
|
View larger version (139K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 18. Focal fatty infiltration. CT scan shows a hypoattenuating lesion (arrow) adjacent to the falciform ligament. This is a typical location for focal fatty infiltration, which should not be mistaken for a metastasis.
|
|
In addition, splenic enhancement is very variable and can be quite heterogeneous, particularly on early contrast-enhanced images. This can appear nodular at times and should not be confused with metastatic disease (Fig 19) (4). Delayed scanning demonstrates homogeneous enhancement of the parenchyma and thus confirms the appearance as perfusion related. The crus of the diaphragm can appear nodular and be mistaken for lymphadenopathy (Fig 20). Conversely, retrocrural lymphadenopathy can be mistaken for a thickened diaphragmatic crus (Figs 21, 22). This can be avoided by establishing the contiguous connection between the crus and adjacent lymph nodes that are separate from this structure.
View larger version (129K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 19a. Nodular splenic enhancement. (a) CT scan shows normal heterogeneous enhancement of the spleen. A nodular appearance of this enhancement (arrow) should not be mistaken for metastatic disease. (b) Delayed image shows homogeneous enhancement of the splenic parenchyma (arrow), an appearance that confirms the presence of a pseudolesion.
|
|
View larger version (138K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 19b. Nodular splenic enhancement. (a) CT scan shows normal heterogeneous enhancement of the spleen. A nodular appearance of this enhancement (arrow) should not be mistaken for metastatic disease. (b) Delayed image shows homogeneous enhancement of the splenic parenchyma (arrow), an appearance that confirms the presence of a pseudolesion.
|
|
View larger version (142K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 21. Retrocrural metastasis in a 55-year-old woman with ovarian carcinoma. CT scan shows a retrocrural metastasis (arrow), which is clearly depicted due to its differential enhancement from that of the diaphragmatic crus (arrowhead).
|
|
View larger version (148K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 22a. Retrocaval lymphadenopathy in a 47-year-old man with lymphoma. (a) On a CT scan, it is difficult to differentiate retrocaval lymphadenopathy (arrow) from the diaphragmatic crus. (b) Image obtained at 6-month follow-up shows some mass effect on the inferior vena cava (arrowhead), a finding that allows differentiation of the lymphadenopathy (arrow) from the normal anatomic structure.
|
|
View larger version (139K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 22b. Retrocaval lymphadenopathy in a 47-year-old man with lymphoma. (a) On a CT scan, it is difficult to differentiate retrocaval lymphadenopathy (arrow) from the diaphragmatic crus. (b) Image obtained at 6-month follow-up shows some mass effect on the inferior vena cava (arrowhead), a finding that allows differentiation of the lymphadenopathy (arrow) from the normal anatomic structure.
|
|
Another pitfall that is encountered in the era of greater anatomic coverage during image acquisition is inclusion of additional anatomic regions such as the low neck in chest CT scans. Lymphadenopathy in the neck may simulate muscles when only partially imaged. In addition, with the increasing use of multirow scanners, CT technologists may include greater anatomic area in the topograms, only to exclude portions when obtaining the diagnostic images. It is not unusual for lesions to be visible on the initial scout images but then excluded from the subsequent diagnostic sequences.
Lymphadenopathy.—
The evaluation of lymphadenopathy is the single most common source of interreader variability (24). Criteria for the evaluation of lymphadenopathy are controversial, and varying assessments of lymph node status are the most common sources of differences in interpretation of oncologic CT scans, accounting for 52% of studies in which a discrepancy was found (4). Of these, 19% of discrepancies were in the mediastinum. Historically, a short-axis diameter of 1 cm has been used as the cutoff; however, recent studies have shown that different size criteria should be applied to different anatomic regions (24,25), with a short-axis diameter as small as 4 mm in the retroperitoneum (24). Guidelines for determination of lymph node enlargement are helpful to avoid reader variability.
|
Combinations of Errors
|
|---|
The preceding categories of errors may combine to produce interpretive errors. Moreover, adverse outcomes may result from additional factors, such as inconsistent measuring and recording of data, failure to communicate important new findings or changes, and lack of familiarity with methods and criteria for measuring tumors.
|
Conclusions
|
|---|
Errors in oncologic CT can be reduced by means of standardized imaging protocols that use intravenous and oral contrast material. Attention should be paid to optimal imaging techniques. Images should be analyzed systematically, with different window settings, and particular attention should be paid to known problem areas and pitfalls according to the underlying disease.
|
Footnotes
|
|---|
Abbreviations: PACS = picture archiving and communication system
|
References
|
|---|
- Hopper KD, Singapuri K, Finkel A. Body CT and oncologic imaging. Radiology 2000;215:27–40.[Abstract/Free Full Text]
- Gollub MJ, Panicek DM, Bach AM, Penalver A, Castellino RA. Clinical importance of reinterpretation of body CT scans obtained elsewhere in patients referred for care at a tertiary cancer center. Radiology 1999;210:109–112.[Abstract/Free Full Text]
- Bechtold RE, Chen MY, Ott DJ, et al. Interpretation of abdominal CT: analysis of errors and their causes. J Comput Assist Tomogr 1997;21:681–685.[CrossRef][Medline]
- Loughrey GJ, Carrington BM, Anderson H, Dobson MJ, Lo Ying Ping F. The value of specialist oncological radiology review of cross-sectional imaging. Clin Radiol 1999;54:149–154.[CrossRef][Medline]
- Hricak H, Kalbhen CL, Scheidler JE, et al. Value of expert interpretation in abdominal oncologic imaging: a multicenter study [abstr]. Radiology 1997; 205(P):225.
- Reason J. Human error: models and management. BMJ 2000;320:768–770.[Free Full Text]
- Lacomis JM, Baron RL, Oliver JH, Nalesnik MA, Federle MP. Cholangiocarcinoma: delayed CT contrast enhancement pattern. Radiology 1997; 203:98–104.[Abstract/Free Full Text]
- Uppot RN, Sahani DV, Hahn PF, Gervais D, Mueller PR. Impact of obesity on medical imaging and image-guided intervention. AJR Am J Roentgenol 2007;188:433–440.[Abstract/Free Full Text]
- Poletti PA, Platon A, Rutschman OT, Schmidlin FR, Iselin CE, Becker CD. Low-dose versus standard-dose CT protocol in patients with suspected renal colic. AJR Am J Roentgenol 2007;188: 927–933.[Abstract/Free Full Text]
- Mulkens TH, Daineffe S, De Wijngaert R, et al. Urinary stone disease: comparison of standard dose and low dose with 4D MDCT tube modulation. AJR Am J Roentgenol 2007;188:553–562.[Abstract/Free Full Text]
- Sahani DV, Kalva SP, Hahn PF, Saini S. 16-MDCT angiography in living kidney donors at various tube potentials: impact on image quality and radiation dose. AJR Am J Roentgenol 2007;188:115–120.[Abstract/Free Full Text]
- Gierada DS, Pilgram TK, Whiting BR, et al. Comparison of standard- and low-dose radiation CT for quantification of emphysema. AJR Am J Roentgenol 2007;188:42–47.[Abstract/Free Full Text]
- Zamboni GA, Gourtsoyianni S, Sourlas E, Raptopoulos V. Value of customized scan timing determined by tracking liver enhancement in oncology patients. J Comput Assist Tomogr (in press).
- Sheafor DH, Frederick MG, Paulson EK, Keogan MT, Dulong DM, Nelson RC. Comparison of un-enhanced, hepatic arterial dominant and portal-venous dominant phase helical CT for the detection of liver metastases in women with breast cancer. AJR Am J Roentgenol 1999;172:961–968.[Abstract/Free Full Text]
- Thompson SE, Raptopoulos V, Sheiman RL, Mc-Nicholas MM, Prassopoulos P. Abdominal helical CT: milk as a low-attenuation oral contrast agent. Radiology 1999;211:870–875.[Abstract/Free Full Text]
- Megibow AJ, Babb JS, Hecht EM, et al. Evaluation of bowel distention and bowel wall appearance by using neutral oral contrast agent for multi–detector row CT. Radiology 2006;238:87–95.[CrossRef][Medline]
- Albertyn LE. Extending the search pattern in computed tomography. Clin Radiol 1987;38:145–149.[CrossRef][Medline]
- Harisinghani MG, Blake MA, Saksena M, et al. Importance and effects of altered workplace ergonomics in modern radiology suites. RadioGraphics 2004;24:615–627.[Abstract/Free Full Text]
- Samuel S, Kundel HL, Nodine CF, Toto LC. Mechanism of satisfaction of search: eye position recordings in the reading of chest radiographs. Radiology 1995;194:895–902.[Abstract/Free Full Text]
- Young ST, Paulson EK, Washington K, Gulliver DJ, Vredenburgh JJ, Baker ME. CT of the liver in patients with metastatic breast carcinoma treated by chemotherapy: findings simulating cirrhosis. AJR Am J Roentgenol 1994;163:1385–1388.[Abstract/Free Full Text]
- Lim HS, Jeong YY, Kang HK, Kim JK, Park JG. Imaging features of hepatocellular carcinoma after transcatheter arterial chemoembolization and radio-frequency ablation. AJR Am J Roentgenol 2006;187:W341–W349.[Abstract/Free Full Text]
- Shirkhoda A. Diagnostic pitfalls in abdominal CT. RadioGraphics 1991;11:969–1002.[Abstract]
- Herts BR, Einstein DM, Paushter DM. Spiral CT of the abdomen: artifacts and potential pitfalls. AJR Am J Roentgenol 1993;161:1185–1190.[Abstract/Free Full Text]
- Hilton S, Herr HW, Teitcher JB, Begg CB, Castellino RA. CT detection of retroperitoneal lymph node metastases in patients with clinical stage I testicular nonseminomatous germ cell cancer: assessment of size and distribution criteria. AJR Am J Roentgenol 1997;169:521–525.[Abstract/Free Full Text]
- Glazer GM, Gross BH, Quint LE, Francis IR, Bookstein FL, Orringer MB. Normal mediastinal lymph nodes: number and size according to American Society mapping. AJR Am J Roentgenol 1985; 144:261–265.[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
S. Mahgerefteh, J. B. Kruskal, C. S. Yam, A. Blachar, and J. Sosna
Peer Review in Diagnostic Radiology: Current State and a Vision for the Future
RadioGraphics,
June 29, 2009;
(2009)
295095086.
[Abstract]
[Full Text]
|
|
|
Top
Abstract
Introduction
