How to Interpret the Data Sets

Beth G. McFarland

CONTENTS

7.1 Introduction 73

7.2 Current and Future Image Display Techniques 73

7.2.1 2D Multiplanar Reformation (2D MPR) 74

7.2.2 3D Endoscopic Fly-Through 75

7.2.3 3D Transparency View (Edge-Enhanced View) 76

7.2.4 Future Advances in Image Display Techniques 76

7.3 Different Categories of Colorectal Morphologies 76

7.3.1 Focal Polypoid Lesions (r/o stool) 76

7.3.2 Pedunculated Lesions 79

7.3.3 Sessile/Flat Lesions

(r/o thick or confluent Folds) 79

7.3.4 Advanced Mural Lesions (r/o collapse) 81

7.4 Standardization of Reporting of Clinically Significant Colorectal Findings 82 References 85

Introduction

CT colonography (CTC) image acquisition and image display capabilities continue to evolve. Concordant with such changes is the need to re-evaluate how to interpret the data. From 2002 to 2005, validation studies have reported a range of results in different patient cohorts using different methods. Important differences in acquisition methods have varied in stool tagging and CT dose techniques, while image display techniques have ranged from primary use of 2D multiplanar reformation to 3D endoscopic fly-through techniques. Specifically the largest cohort to date of 1233 patients reported excellent results using rigorous tagging and electronic subtraction, with reader evaluation of primary 3D fly-through review (Pickhardt et al. 2003). Three other multi-

Diagnostic Imaging Associates, Center for Diagnostic Imaging, St. Luke's Hospital, 232 S. Woods Mill Rd, Chesterfield, MO 63017, USA and Adjunct Professor, Washington University School of Medicine, Mallinckrodt Institute of Radiology, 510 S. Kingshighway Blvd, St. Louis, MO 63110, USA

institutional studies reported less good results using no tagging or subtraction and predominant use of 2D MPR as primary review (Cotton et al. 2004; Johnson et al. 2003; Rockey et al. 2005). Recently a series of200 patients with excellent results used stool tagging without catharsis or subtraction, exploiting ultra low dose techniques and 2D MPR as a primary review (Iannaccone et al. 2004). Certainly the differences in these studies are complex; however key influences of image display techniques for data interpretation are important to understand.

This chapter will focus on 1) current and future image display techniques for data interpretation, 2) application of these techniques in the major categories of colorectal morphologies, and 3) issues of standardization of reporting clinically significant colorectal findings in CTC (C-RADS).

Current and Future Image Display Techniques

After years of diligent use of 2D multiplanar reformation (2D MPR) as a primary review with 3D to problem solve, the success of the study of PickhardT et al. (Pickhardt et al. 2003), aided by improved computer graphics, eclipsed the field in 2003 to demonstrate that 3D as a primary review was not only feasible but may be better. Unfortunately, there has been somewhat of a binary debate of whether 2D vs 3D is better, rather than an understanding of how to apply each of these techniques cohesively in the appropriate setting.

How are we now to approach data interpretation? Currently, the answer probably is a seamless interaction between 2D MPR and 3D endoscopic fly-through techniques, which may vary across patients or within specific colonic segments of a patient. The 3D fly-through as a primary review uses 3D to detect, with 2D MPR to help characterize. Conversely, use of 2D MPR as a primary review uses 2D to detect, with 3D MPR to help characterize. In addition to these techniques, the 3D transparency view allows an overall view of the colonic anatomy, simulating a barium enema visualization. In order to apply optimally techniques best, it is important to understand the uses, advantages and disadvantages.

2D Multiplanar Reformation (2D MPR)

Primary 2D MPR review for lesion detection can provide a time efficient evaluation of the colon, exploiting an extra-colonic field of view for improved orientation. This visualization is based on a real time sectoring through the colonic sub-segments in cine mode, in a continuous direction typically from the rectum to the cecum. Two window levels settings are typically utilized. The polyp window (width 1500, level -200) imparts the high contrast interface to detect intraluminal colorectal polypoid lesions. The soft tissue window setting (width 400, level 10) is critical to evaluate more advanced wall lesions, discern the high density of false positives or the fat density of lipomas, and to evaluate the extra-colonic findings. In some areas of retained dense fluid, a narrower window setting may be needed to better see within the fluid level.

When a lesion is detected in the polyp window of 2D MPR, the following steps can be taken to most efficiently characterize a lesion. First, the density of the lesion in the soft tissue window is directly available. If the lesion is dense (i.e., false positive of stool) or fatty (lipoma), typically no further evaluation is needed. If however the lesion is soft tissue in density, both the 2D MPR additional views (e.g., sagittal and coronal) and the 3D endoscopic view can then be used to assess morphology of a focal finding better. The use of 3D to further characterize can be very helpful to evaluate the common dilemma of discernment of a thickened fold (or confluence of folds) from a polyp on a fold. Lastly, confirmation of a lesion's anatomic location on the corresponding prone (or supine) data set is needed. Although automated registration of supine and prone data sets is becoming technically feasible, the in vivo range of change in position of a focal finding between prone and supine positions might make 2D MPR with its extra-colonic field of view most efficient to localize a given area. Confirmation of the same location on both prone and supine data sets (when image quality permits) greatly increases confidence of a true positive lesion.

To use 2D MPR as a primary review, it can be helpful to perform initially a one minute coronal cine of both the prone and supine data sets to get the "lay of the land". This allows the reader to determine the degree of tortuosity and the colonic course, as well as the image quality of colonic distention and fluid/ stool retention. The data set with the best image quality can be chosen to evaluate first. Typically, as each lesion is detected, characterization of the lesion on the other 2D MPR views and 3D views can be made between prone and supine data sets. Images of true positive lesions are taken and an internet report of findings for future dictation can be started. This type of organization of effort, involving lesion detection and characterization, followed by image capture of important findings, along the continuous retrograde path of the colon from rectum to cecum, allows an efficient and thorough evaluation of the colonic surface area. If interrupted by another task, the colon segment reached can be recorded and then evaluation can be reinitiated at this segment, once the interpretation can be resumed. After finishing the first data set (typically supine) with characterization of each lesion detected, the other data set (typically prone) can be briefly viewed with an axial cine to see potentially any additional findings. Using embedded arrows for lesions already detected prevents one wasting time re-evaluating the same lesions between data sets.

What are the advantages and disadvantages of 2D MPR? The use of 2D MPR offers several important advantages. First there is the direct display of the source attenuation data of a focal lesion to determine density. Specifically, the density of a focal lesion, such as fat in a lipoma or high density or focal pockets of air in stool, provides the most important characteristics to confirm false positive lesions. Another advantage of 2D MPR is the ability to visualize the location of a lesion from an extralu-minal viewpoint, rather than the immersed endo-scopic view of 3D fly-through. This can be helpful to confirm whether a lesion is in the same segment between prone and supine data sets (especially helpful in tortuous colons), as well as whether the lesion is dependent or non-dependent within a given segment. If a lesion shifts in position to different segments or changes to a dependent position on both prone and supine views without a visualized stalk, the concern for a false positive (e.g., retained stool) increases. Another advantage is that 2D MPR can be a time efficient evaluation of the colonic findings, since a thorough cine of both data sets is done once, compared to the need of forward and retrograde 3D

fly-through views in both prone and supine data sets. Subtle mural lesions can often best be seen in the MPR soft tissue window settings. The disadvantages of 2D MPR as a primary review may be decreased sensitivity compared to the increased surface area visualization of 3D fly-through. Although this has not been directly confirmed in validation studies, the improved results of 3D as a primary review in the Pickhardt et al. study are compelling. Reader fatigue can also be greater with 2D MPR, given the potentially more subtle and briefer visualization of lesions during the cine method of sectoring through the data.

3D Endoscopic Fly-Through

The use of 3D fly-through as a primary review provides a continuous fly-through of the colon, using the endoscopic field of view. The exciting advances of this technique have become more generalizable across 3D workstations; however differences still exist. In some 3D workstations, preprocessing of the data with segmentation (i.e., selective extraction) of the colon is first done, followed by calculation of the central path through the entire colon. Other workstations do not segment the colon or calculate the central path. For the reader, the review typically begins within the rectum, with continuous retrograde fly-through to the cecum.

When a focal lesion is found in 3D, further characterization can be done in several ways. First the 3D morphology is initially directly viewed, which can efficiently display overall anatomy and relationship of a lesion to surrounding folds. Second, assessment of density can then be performed with simultaneous registration of the finding in 2D MPR to display directly the density of the lesion. Some vendors provide an opacity map of the lesion within the 3D endoscopic view, which helps display the profile of density differences across the lesion. Finally, confirmation of the position and anatomic location of the lesion in both supine and prone data sets is again needed to increase confidence of a true positive lesion. This may best be performed with comparison of the lesion between MPR views for extraluminal orientation.

To use 3D as a primary review, endoscopic fly-through in antegrade and retrograde paths for both supine and prone data sets are necessary. Some advocates of 3D however state that if excellent visualization with no lesions found is present after for ward and backward fly-through paths of the supine data set, only a retrograde fly-through is needed in the prone data set (e.g., eliminates antegrade fly-through in the second data set). Similar to 2D MPR, use of embedded arrows in lesions already evaluated keeps evaluation of additional lesions efficient between prone and supine data sets. A critical point is the need to also sector through the axial MPR data in soft tissue settings (done at start or end of review) to exclude any subtle or advanced mural lesion. Circumferential narrowing or partial wall involvement of the colon in advanced cancers can be more subtle in the immersed 3D endoscopic view, whereas soft tissue wall thickening or irregularity can be better seen in the extraluminal viewpoint of 2D MPR.

What are the advantages and disadvantages of 3D endoscopic fly-through techniques? A strong advantage is the increased surface area visualization, which continues to be aided by improved navigational tools. In non-tortuous segments of colon, there is a longer period of visualization of a focal colorectal lesion over the course of the fly-through, compared to the brief visualization seen while sectoring through a sub segment of the colon using 2D MPR cine techniques. This advantage however is diminished in marked areas of tortuosity or areas of collapse. Also focal polypoid lesions can be more visibly apparent within the colon lumen, compared to 2D, and thus can be easier to see in 3D fly-through. Both the potential advantages of longer visualization and increased ease of visualization of a lesion can lead to greater detection rates with less reader fatigue. The disadvantages of 3D endoscopic fly-through can include longer length of evaluation to complete review of the antegrade and retrograde paths. In tortuous colons, surface areas visualization around the inner curve of a turn can be initially missed. Importantly, any increased sensitivity to detect more lesions needs to be scrutinized relative to specificity. Namely, improvements in lesion detection with increased surface area visualization must be achieved, without increased false positive rates.

In summary, use of 2D MPR and 3D display techniques optimally are best used with seamless integration to exploit their inherent advantages and diminish their disadvantages. Although there is increasing use of the 3D fly-through technique with improved computer graphics and navigational aids, use of 2D MPR as a primary review may still be needed in specific segments or sub-segments not amenable to 3D. Specifically, areas best evaluated with 2D MPR primarily may include areas of marked muscular hypertrophy from diverticulosis, sharp hairpin turns, colonic collapse, or marked fluid retention. In these areas, the 3D fly-through may be suspended, with transition to 2D MPR through a given region. In addition, in areas where multiple focal findings are being detected in 3D raising the concern for stool retention, evaluation in 2D MPR may allow a better overall characterization. Thus, given differences in image quality and anatomy which vary within or between patients, complementary use of 2D or 3D can be selectively utilized for improved visualization.

Appropriate training is important to acquire the new skills of these techniques. Currently both academic and commercial programs for reader training are available. A highly recommended source of these sites is available at the International Symposium of Virtual Colonoscopy held each fall in Boston. With training, primary review of the colorectal lesions (not including extra-colonic findings) using either technique to complete a normal data set with good to excellent image quality has been stated by experts to range from five to ten minutes. A recently developed clinical service may benefit initially from double reading of cases among trained colleagues.

3D Transparency View (Edge-Enhanced View)

In addition to the primary modes of interpretation, the 3D transparency view provides an effective visualization to display the overall colonic anatomy and to demarcate where focal findings are. Although this view is not helpful in making the diagnosis, its role to summarize the findings in a consistent and accurate way is important for current management and future surveillance of lesions. Similar to the barium enema in appearance, this view can give an effective roadmap for the gastroenterologist or surgeon. Even in cases with no lesions, standard AP and bilateral oblique views can be helpful for future reference.

Future Advances in Image Display Techniques

How we interpret virtual colonography in 2005 will probably change dramatically in the next five to ten years. Important influences will be the further refinement of computed aided diagnosis and molecular imaging. Computed aided diagnostic algorithms in CTC are being actively explored in academic and commercial efforts. If sensitivity can be achieved across the complexity of colorectal morphologies, without a compromise of specificity, remarkable efficiency of the data evaluation will be achieved. Whether CAD is used optimally as a primary read or a secondary read will be interesting to evaluate. As molecular imaging techniques continue to evaluate functional information at cellular and molecular levels, colorectal applications may shift to other modalities, such as MR and optical imaging. Certainly the success of molecular imaging could lead to a phenomenal break-through of detection of clinically significant lesions in the polyp-carcinoma pathological continuum, along with focused therapy.

Different Categories of Colorectal Morphologies

There are common types of colorectal morphologies evaluated in CT colonography. These include the focal polypoid lesion, pedunculated lesion, flat or sessile lesion and advanced mural lesions. This section will describe these morphologies and their corresponding false positive counterparts. The differential application of 2D and 3D image displays to assess these morphologies will also be reinforced.

Focal Polypoid Lesions (r/o stool)

One of the most common colorectal morphologies is the focal polypoid lesion. This is also the most common morphology of the false positive lesion of retained stool. Thus discernment between a focal polyp and stool are critical. Key features include the following:

The morphology of a focal polyp is typically smooth and round (Fig. 7.1). Although stool can also be similar in morphology, margins which are more geometrical or angular are highly suggestive of stool (Fig. 7.2). A polypoid lesion is typically soft tissue in density; however lesions which are increased in density are highly specific of stool. Stool can also be low attenuation or opaque; however this can overlap with the partial volume effects of smaller polyps, depending on the collimation used. A focal polyp can have air around the edges where it abuts the wall; however central focal pockets of air within a lesion

Fig. 7.1a-e. Typical features of a true positive polypoid morphology (arrows), demonstrating smooth margins, soft tissue density and constant location between prone and supine positions: a prone axial 2D MPR and b supine axial 2D MPR in polyp settings (W 1500, L -200); c axial 2D MPR with soft tissue settings (W 400, L 10) confirms polyp density; d 3D perspective volume rendered view demonstrates smooth polypoid morphology; e 3D transparency view gives roadmap of colonic anatomy, with arrow demarcating location of lesion

Fig. 7.1a-e. Typical features of a true positive polypoid morphology (arrows), demonstrating smooth margins, soft tissue density and constant location between prone and supine positions: a prone axial 2D MPR and b supine axial 2D MPR in polyp settings (W 1500, L -200); c axial 2D MPR with soft tissue settings (W 400, L 10) confirms polyp density; d 3D perspective volume rendered view demonstrates smooth polypoid morphology; e 3D transparency view gives roadmap of colonic anatomy, with arrow demarcating location of lesion is diagnostic of stool (Fig. 7.3). Shift vs constancy of location of a focal finding relative to the colon wall is important. A polypoid lesion will stay fixed in the same position relative to the wall between prone and supine images. In contrast, the finding of stool which drops dependently is highly characteristic (Fig. 7.4). There are exceptions to this. A sigmoid polyp may appear to shift dependently on prone and supine images, but actually the redundancy of the sigmoid mesentery is what shifts (Laks et al. 2004). Stool can be adherent to the wall, which has been described with the phospha-soda bowel preparations. Also, a pedunculated polyp which does not demonstrate its stalk can appear to drop dependently. Finally, if intravenous contrast is used, differences in enhancement can be seen. Polyps potentially can enhance, w R

Fig. 7.2a-d. Typical features of false positive retained stool (arrows), demonstrating high density or angular margins: a axial 2D MPR (W 1500, L -200) shows focal polypoid lesion, compared to b; b axial 2D MPR (W 400, L 10) in soft tissue settings better demonstrates high density of stool; c axial 2D MPR view shows angular margins of another area of stool; d 3D perspective volume rendered view demonstrates dilemma of detection of multiple focal lesions, some demonstrating angular margins of stool

Fig. 7.2a-d. Typical features of false positive retained stool (arrows), demonstrating high density or angular margins: a axial 2D MPR (W 1500, L -200) shows focal polypoid lesion, compared to b; b axial 2D MPR (W 400, L 10) in soft tissue settings better demonstrates high density of stool; c axial 2D MPR view shows angular margins of another area of stool; d 3D perspective volume rendered view demonstrates dilemma of detection of multiple focal lesions, some demonstrating angular margins of stool

Fig. 7.3a,b. Polypoid lesions with focal pockets of air seen in true polyp vs stool, best shown in axial 2D MPR: a true positive sessile polyp (arrow) with air around edges of lesion (arrowheads), where lesion abuts the wall; b false positive of stool (arrow) with central pockets of air (arrowheads)

whereas stool will not enhance. As diagnostic CT is exploited to stage the liver along with evaluation of the colorectal lesions, the increasing use of IV contrast may lead to better understanding of the enhancement characteristics of polyps.

Both 2D and 3D techniques help to evaluate these characteristics. The morphologic features of round or smooth vs angular or geometric margins are best seen with 3D endoscopic views. Inherent density, location, focal pockets of air, and degree of enhance

Fig. 7.4a-d. False positive lesion of retained pill (arrow) with characteristic shift of position: a axial 2D MPR supine image (W 1500, L -200) demonstrates a polypoid lesion; b axial 2D MPR view in soft tissue window settings (W 400, L 10) better shows low density of pill; c axial 2D MRP prone image in soft tissue settings better demonstrates shift to dependent position of pill, consistent with false positive; d 3D volume rendered view of pill mimics a polyp

Fig. 7.4a-d. False positive lesion of retained pill (arrow) with characteristic shift of position: a axial 2D MPR supine image (W 1500, L -200) demonstrates a polypoid lesion; b axial 2D MPR view in soft tissue window settings (W 400, L 10) better shows low density of pill; c axial 2D MRP prone image in soft tissue settings better demonstrates shift to dependent position of pill, consistent with false positive; d 3D volume rendered view of pill mimics a polyp ment is best evaluated with 2D MPR. As stool tagging continues to be developed, the ability to see high density within residual stool will further aid differentiation of polyp from stool.

Pedunculated Lesions

The pedunculated lesion, comprised of a round polyp head with a linear stalk, is a very distinguishable lesion. The challenge is how different these lesions can appear between prone and supine data sets (Fig. 7.5). On one data set, the polyp head can be suspended dependently from the stalk, whereas on the corresponding data set both polyp head and stalk can lie dependently together along the wall. If the stalk is long enough, lesions on the border between two segments can lie in different adjacent segments between prone and supine views. Also of importance is the variability in lesion measurement with such morphological differences between views.

A consistent reporting style is to measure the polyp head, with exclusion of the stalk. Thus, choosing the image where the polyp head and stalk are best discerned provides the more accurate and reproducible measurement (Fig. 7.5).

In general, 3D endoscopic views can provide improved visualization of these morphological features. One exception would be the visualization of the highly characteristic stalk of a pendunculated lesion in a segment with marked muscular hypertrophy of diverticulosis. In this setting, 2D MPR may offer an advantage, due to the impaired endoscopic visualization within the thickened folds (Fig. 7.6).

Sessile/Flat Lesions

(r/o thick or confluent Folds)

These lesions are currently considered the most challenging lesions to detect (Fig. 7.7). One important issue is the variability in definition of the terms.

Fig. 7.5a-d. Typical features of a pedunculated polyp in sigmoid colon, demonstrating polyp head (arrow) and characteristic stalk (arrowheads), with marked changes between positions: a axial 2D MPR and b 3D volume rendered view in prone position demonstrates polyp head and stalk. Accurate measurement of long axis of polyp head, excluding the stalk, is shown; c axial 2D MPR and d 3D volume rendered view in prone position shows shift to dependent position with corresponding change in morphology

Fig. 7.5a-d. Typical features of a pedunculated polyp in sigmoid colon, demonstrating polyp head (arrow) and characteristic stalk (arrowheads), with marked changes between positions: a axial 2D MPR and b 3D volume rendered view in prone position demonstrates polyp head and stalk. Accurate measurement of long axis of polyp head, excluding the stalk, is shown; c axial 2D MPR and d 3D volume rendered view in prone position shows shift to dependent position with corresponding change in morphology

Fig. 7.6a,b. Pedunculated lesion in area of luminal narrowing and marked diverticulosis: a axial 2D MPR view best visualizes the characteristic stalk (arrows) from the polyp head (arrowhead), compared to b; b 3D volume rendered view, with polyp head shown (arrowheads), but stalk obscured by luminal narrowing and muscular hypertrophy

Fig. 7.6a,b. Pedunculated lesion in area of luminal narrowing and marked diverticulosis: a axial 2D MPR view best visualizes the characteristic stalk (arrows) from the polyp head (arrowhead), compared to b; b 3D volume rendered view, with polyp head shown (arrowheads), but stalk obscured by luminal narrowing and muscular hypertrophy a c

Fig. 7.7a-d. Sessile lesion (arrows) along a fold poses a challenge in detection: a,b axial 2D MPR images partially demonstrate the sessile lesion as an asymmetric bulge along the fold (best demonstrated with cine motion not shown); c 3D volume rendered view better demonstrates overall morphology of lesion; d correlative view from optical colonoscopy (copyright given by Radiology for MoFarland et al. 2002)

Fig. 7.7a-d. Sessile lesion (arrows) along a fold poses a challenge in detection: a,b axial 2D MPR images partially demonstrate the sessile lesion as an asymmetric bulge along the fold (best demonstrated with cine motion not shown); c 3D volume rendered view better demonstrates overall morphology of lesion; d correlative view from optical colonoscopy (copyright given by Radiology for MoFarland et al. 2002)

"Sessile" is generally defined as a lesion with a polyp base that is twice as long as the polyp height. Thus a 1 cm high lesion with a 2-cm base could be defined as sessile. This is very different from a "flat" lesion which is generally defined as a lesion measuring 1-3 mm in height. Unfortunately, the literature is inconsistent with these definitions and reported sensitivities have varied. Fidler et al. (2002) reported a sensitivity range of 15-65% to detect 22 sessile polyps (size range of 0.4-3.5 cm) in a cohort of 547 patients. Pickhardt et al. (2004) reported a sensitivity of 80% (47/59 polyps) to detect sessile lesions (defined as lesions with a base greater than height) in the cohort of 1233 screening patients.

The most common false positive counterpart to the flat or sessile lesion is a thickened fold or conflu ence of folds (Fig. 7.8). For these types of challenging lesions, both 2D MPR (in soft tissue window-level settings) and 3D views have advantages and disadvantages. Both techniques should be explored to detect and characterize these lesions best.

Advanced Mural Lesions (r/o collapse)

CT colonography has reliably shown high sensitivity to detect advanced mural lesions (Fig. 7.9). A potential challenge in CT colonography is the discernment between an advanced mural lesion of advanced cancer from an area of focal collapse with relative wall thickening (often seen at points of flex b a c a c

Fig. 7.8a-c. False positive of confluence of folds (arrows) mimicking a flat lesion: a 3D volume rendered view raises concern for a flat lesion; b axial 2D MPR in soft tissue settings also demonstrates flat morphology; c sagittal 2D MPR best demonstrates confluence of folds

Fig. 7.8a-c. False positive of confluence of folds (arrows) mimicking a flat lesion: a 3D volume rendered view raises concern for a flat lesion; b axial 2D MPR in soft tissue settings also demonstrates flat morphology; c sagittal 2D MPR best demonstrates confluence of folds b a c ures, tortuosity, or muscular hypertrophy). At CT colonography, an advanced cancer will tend to stay fixed between prone and supine imaging, whereas an area of collapse can change in distention between positions. A cancer will typically have irregularity of the soft tissue rind along the wall, whereas focal collapse or muscular hypertrophy will be more smooth and symmetric (Fig. 7.10). If intravenous contrast is given, a cancer may have more irregular enhancement, whereas focal collapse of normal colon will enhance symmetrically.

The importance of 2D MPR with soft tissue settings (window 400, level 10) needs to be emphasized with these types of lesions. Whether this is subtle mural thickening or advanced, the 2D MPR views give valuable information of the mural relationships, which extend beyond the "lumenography" of the 3D fly-through (Fig. 7.11). In addition the 3D transparency view, which simulates the barium enema, can be a powerful view to display the lesion for others to appreciate the size and location of the cancer.

Standardization of Reporting of Clinically Significant Colorectal Findings

The Virtual Colonoscopy Working Group, represented by an organized group of CTC investigators at the International Virtual Colonoscopy annual meeting, recently published a consensus statement regarding the early development of a reporting structure of colorectal findings (Zalis et al. 2005). This reporting structure, named "C-RADS" helps to lay the groundwork for structured reporting of lesion morphology, size, and location, and preliminary recommendations of lesion surveillance. Although this effort now needs to be refined with multi-disciplinary consensus, it represents an important step towards more consistent and reproducible reporting at CT colonography.

C-RADS describes the use of three morphologies of lesions: sessile (broad based lesion whose width is greater than its vertical height), pedunculated (polyp with a separate stalk), and flat (polyp with vertical

Fig. 7.9a-d. Advanced mural lesion (arrows): a prone non-contrast axial 2D MPR demonstrates advanced mural lesion; b supine contrast enhanced axial 2D MPR shows enhancing mass immersed in fluid; c 3D volume rendered intraluminal view demonstrates mural mass of cancer; d 3D transparency view shows classic apple-core lesion

Fig. 7.9a-d. Advanced mural lesion (arrows): a prone non-contrast axial 2D MPR demonstrates advanced mural lesion; b supine contrast enhanced axial 2D MPR shows enhancing mass immersed in fluid; c 3D volume rendered intraluminal view demonstrates mural mass of cancer; d 3D transparency view shows classic apple-core lesion

Vulvar Lesion
Fig. 7.10a,b. Advanced mural lesion (arrows) vs collapse (arrowheads): a sagittal 2D MPR view shows mural thickening of advanced cancer vs area of luminal collapse without mural thickening; b 3D edge enhanced view also demonstrates these two areas

Fig. 7.11a-e. Subtle advanced mural lesion with polypoid (white arrow) and stalk (white arrowheads) components, along with infiltrative T3 mural invasion (open arrows), best seen in soft tissue MPR views: a supine axial 2D MPR (W 1500, L -200) does not demonstrate lesion well, compared to b; b supine axial 2D MPR in soft tissue settings (W 400, L 20) best demonstrates the polypoid component and flat soft tissue mural infiltration; c prone axial 2D MPR shows immersed lesion requires a narrower soft tissue window setting (W 900, L 300) to see through the fluid; d optimized 3D view shows polypoid and infiltrative mural components (only seen retrospectively); e corresponding view at optical colonoscopy

Fig. 7.11a-e. Subtle advanced mural lesion with polypoid (white arrow) and stalk (white arrowheads) components, along with infiltrative T3 mural invasion (open arrows), best seen in soft tissue MPR views: a supine axial 2D MPR (W 1500, L -200) does not demonstrate lesion well, compared to b; b supine axial 2D MPR in soft tissue settings (W 400, L 20) best demonstrates the polypoid component and flat soft tissue mural infiltration; c prone axial 2D MPR shows immersed lesion requires a narrower soft tissue window setting (W 900, L 300) to see through the fluid; d optimized 3D view shows polypoid and infiltrative mural components (only seen retrospectively); e corresponding view at optical colonoscopy

M d c e height less than 3 mm). Lesion measurement is critical for clinical management and can vary among readers. C-RADS defines the measurement of a polypoid lesion to be the maximal long axis of the polyp head, with exclusion of the stalk if present. For more sessile lesions, the maximal length along the base of the polyp should be used. Both 2D MPR and 3D techniques have been advocated for measurement, both of which certainly are felt to exceed the accuracy of axial measurements (Pickhardt 2005).

Lesion location refers to the standardized colonic segmental divisions of rectum, sigmoid, descending, transverse, ascending and cecum.

Recommendations of lesion surveillance are of active debate. At the heart of this issue is the controversy of what is the clinical index lesion of significance. In a recent future trends report initiated by the American Gastroenterology Association, the clinical significance of the intermediate 6-9 mm polyp poses the largest debate (Van Dam et al. 2004). Further multi-disciplinary consensus will now be needed to refine these recommendations.

The suggested categories of lesions are as follows. A C0 category represents an inadequate study (e.g., inadequate prep or insufflation) or a study awaiting prior comparisons. A C1 category is a normal colon (no lesions or lesions <5 mm) or a benign lesion (e.g., lipoma). A C2 category is an intermediate polyp (e.g., less than three 6- to 9-mm polyps) or indeterminate finding (e.g., cannot exclude a polyp >6 mm). A C3 category designates clinically significant polyps (e.g., polyps >10 mm or greater than three 6- to 9-mm polyps). A C4 category is a colonic mass, which is likely malignant. Surveillance guidelines are still being refined; however each patient needs to be evaluated in the clinical context of age, comorbidity, colorectal symptoms and a priori concern for colorectal cancer.

Finally, C-RADS discusses the reporting of extra-colonic findings. To achieve cost effectiveness, the judicious reporting of extra-colonic findings to minimize unnecessary additional imaging tests will be critical. Significant extra-colonic findings, such as abdominal aortic aneurysms, solid renal or liver masses, adenopathy and lung nodules (greater than 1 cm) are emphasized. Less significant findings, such as small liver and renal cystic lesions (common findings which are often difficult to characterize without contrast), and gallstones hopefully will be under-reported. Further definition of how to categorize and follow these findings will be important.

In summary, C-RADS begins the process of standardization during the continuum of change and improvement. The RSNA initiative of RADLEX will mature the elements of informatics to this process. Multi-disciplinary consensus of radiologists, gastroenterologists, internists, surgeons, patholo-gists, and epidemiologists will hopefully continue to define and standardize clinical issues of importance. At the core of these issues will be definition of what constitutes a clinically significant lesion, with appropriate surveillance recommended within the clinical context of comorbidity, age and colorec-tal symptoms. As standardization is refined during ongoing evolution of the technique, an active process of quality assurance will be helpful for community implementation.

References

Cotton PB, Durkalski VL, Pineau BC, Palesch YY, Mauldin PD, Hoffman B et al. (2004) Computed tomographic colonoscopy (virtual colonoscopy): a multi-center comparison with standard colonoscopy for detection of colorectal neoplasia. JAMA 291:1713-1719 Fidler JL, Johnson CD, MacCarty RL, Welch TJ, Hara AK, Harmsen WS (2002) Detection of flat lesions in the colon with CT colonography. Abdom Imaging 27:292-300 Iannaccone R, Laghi A, Catalono C, Mangiapane F, Lamazza A, Schillaci A et al. (2004) Computed tomographic colonography without cathartic preparation for the detection of colorectal polyps. Gastroenterology 127:1200-1211 Johnson CD, Harmsen WS, Wilson LA, Maccarty RL, Welch TJ, Ilstrup DM et al. (2003) Prospective blinded evaluation of computed tomographic colonography for screen detection of colorectal polyps. Gastroenterology 125:311-319 Laks S, Macari M, Bini E (2004) Positional changes in colon polyps at CT colonography. Radiology 231:761-766 McFarland EG, Pilgram TK, Brink JA, et al. (2002) CT colon-graphy: multi-observer diagnostic performance. Radiology 225:380-390

Pickhardt PJ, Choi JR, Hwang I, Butler JA, Puckett ML, Hildebrandt HA et al. (2003) Computed tomographic virtual colonoscopy to screen for colorectal neoplasia in asymptomatic adults. N Eng J Med 349:2191-2200 Pickhardt PJ, Nugent PA, Choi JR, Schindler WR (2004) Flat colorectal lesions in asymptomatic adults: implications for screening with CT virtual colonoscopy. AJR 183:13431347

Pickhardt PJ, Lee AD, McFarland EG, Taylor AJ (2005) Linear polyp measurement at CT colonography: in vitro and in vivo comparison of two-dimensional and three-dimensional displays. Radiology 236:872-878 Rockey DC, Paulson E, Niedzwiecki D, Davis W, Bosworth HB, Sanders L et al. (2005) Analysis of air contrast barium enema, computed tomographic colonography, and colonoscopy: prospective comparison. Lancet 365(9456):305-311 Van Dam J, Cotton P, Johnson CD, McFarland BG, Pineau BC, Provenzale D, Ransohoff D, Rex D, Rockey D, Wootton T (2004) AGA future trends report: CT colonography. Gas-troenterology 127:970-984 Zalis ME, Barish MA, Choi JR, Dachman AH, Fenlon HM, Ferrucci JT, Glick SN, Laghi A, Macari M, McFarland EG, Morrin MM, Pickhardt PJ, Soto J, Yee J (2005) CT colonog-raphy reporting and data system: a consensus proposal. Radiology 236:3-9

Managing Diverticular Disease

Managing Diverticular Disease

Stop The Pain. Manage Your Diverticular Disease And Live A Pain Free Life. No Pain, No Fear, Full Control Normal Life Again. Diverticular Disease can stop you from doing all the things you love. Seeing friends, playing with the kids... even trying to watch your favorite television shows.

Get My Free Ebook


Post a comment