Lorenzo Cereser, MD

Computer-aided detection (CAD) tools use software to analyze digital or digitized images to find features associated with the target disease. The objective of CAD is to mark suspicious findings on the monitor and/or on a print of the image to help radiologists detect lesions (Figure 1).

 The first CAD tools were approved by the U.S. Food and Drug Administration for clinical use in 1998. Several commercial and noncommercial CAD systems have since become available.

CAD algorithms work by identifying areas of signal in images that may contain a cancer. The standard approach is to use thresholding algorithms to identify as many true signals and as few false signals as possible. Signal data are separated from the background in a process known as segmentation. The signals are then subjected to a probabilistic analysis to assess the likelihood that the structure on the image contains malignancy-induced abnormalities.

 The final result is a CAD prompt if the probability of cancer being present is sufficiently high.

All CAD systems must also be trained on a database of real cases before they are used in clinical practice. These cases should ideally be proven by biopsy or, if not, by follow-up of at least two years. The incidence and type of disease should mirror that expected in the target population.1 A basic method for assessing CAD performance during the training period is to plot the free-response receiver operating characteristic (FROC). This represents the sensitivity of the system in detecting cancer as a function of the averaged number of false-positive marks on each image. The best performing CAD systems will have a low false-positive rate as well as high sensitivity.

CAD tools are well-suited to mammographic screening for breast cancer. Screening has been shown to be a cost-effective way of detecting (and thus treating) breast cancer at an early stage. It is, however, associated with a false-negative rate of up to 25%.

 False negatives are cancers that are overlooked at the baseline examination and either appear as interval cancers or are detected on subsequent screening mammograms. Between 27% and 70% of these cancers turn out to be visible on a retrospective analysis of the baseline mammogram.

The excessive false-negative rate in breast screening has been interpreted as a consequence of the lower prevalence of disease compared with the clinical scenario (fewer and smaller lesions). Thus, small carcinomas detected during screening have less obvious mammographic features than symptomatic lesions. Screening is a highly demanding task involving a detailed visual search for subtle signs.

One solution to the problem of false negatives is to have two radiologists reading mammograms independently, with a consensus review if their opinions on recall differ. This practice of double-reading has been implemented at many institutions in Europe. It can reduce the false-negative rate by 15%, but at a high cost in terms of human and material resources.

An alternative solution is to replace the second reader with CAD. This is based on the principle that software algorithms provide a constant efficiency over time and will allow lesions overlooked by a single reader to be detected reliably. CAD is also cheaper than hiring a second reader.

 This option is more common in U.S. breast clinics and hospitals than in Europe.

The main measure of a CAD system’s performance is its intrinsic sensitivity in detecting cancer. CAD prompts indicate either true cancers (true positives; TP), or regions that the algorithm mistakenly believes to be abnormal (false positives; FP). CAD must typically have a high intrinsic sensitivity (TP/TP+FN) to increase radiologists’ sensitivity. This will lead to an improved cancer detection rate.

The advantage of high sensitivity is balanced by the false-positive rate. CAD tools typically provide several marks per case, most of which are false-positive calls. The higher the sensitivity of the CAD package, the greater the number of prompts on the screen, and the higher the number of false positives.

 This decrease in specificity means that more patients in the screening program could be recalled for further investigation and more needless breast biopsies could be performed. In other words, the positive predictive value of the screening test is expected to decrease. It has been estimated that for every 1% increase in the recall rate, the rate of cancers detected rises by 0.022%.

Intrinsic CAD performance is, then, a compromise between sensitivity and specificity. Any adjustments are usually made in favor of raising the sensitivity.

The analysis of digital radiography and CAD in breast screening programs is incomplete. No randomized controlled trials have been performed to assess changes in survival. The performance measures described above are consequently used as surrogate endpoints to evaluate clinical performance.

The studies published in this area fall into two categories: retrospective and prospective.

 Retrospective studies are designed to evaluate the sensitivity of CAD on previous screening mammograms with a view to checking if the software would have picked up missed cancers. An overview of results (Table 1)

 shows that CAD provides adequate sensitivity in detecting cancer, and that this sensitivity is higher for microcalcifications than for masses. These studies also show that CAD can help radiologists by retrieving missed cancers.

Retrospective studies are, however, biased by the higher prevalence of cancer in study populations compared with the screening setting, and by the incorrect assumption that radiologists always accept prompted suggestions. Contrary to this assumption,

 readers tend to ignore the majority of CAD marks; up to 16% of lesions they detect were missed by the software system (Figure 2).

It is consequently recommended that radiologists use CAD only after the preliminary reading, and that if CAD fails to mark an area the reader thought was suspicious, the patient should still be recalled.

Prospective clinical trials investigating CAD in a screening scenario require larger populations to be studied. These will, however, lead to a more accurate estimation of CAD’s impact on readers and to better predictions of recall rates. Studies can be matched (radiologists read mammograms before and after CAD) or unmatched (periods of clinical practice without CAD and then with CAD are compared). The results tend to show that CAD leads to an increase in both the cancer detection rate and the recall rate (

The clinical value of CAD remains a topic for debate. Some trials and meta-analyses suggest that CAD is not useful at all.

 It is difficult to get an overall picture, though, given the wide variation in the design and purpose of different studies. For example, the relative proportions of masses and microcalcifications in the study population may vary, the background density may make it easier or harder to see masses, CAD technology may have evolved from one study to the next, different reading protocols may be employed, etc.

It should not be forgotten that reader experience can affect CAD performance considerably. CAD is generally assumed to be of more help to inexperienced radiologists. This is because practitioners with less experience making diagnostic decisions will be more likely to be influenced by a positive or negative CAD result.

Any evaluation of the influence of CAD on readers should consider the complex cognitive features related to human perception. Newer CAD systems may produce fewer prompts, but readers will still be confronted with a range of marks to examine in detail, most being false positives. Radiologists may develop unconscious strategies to reduce the effort of scrutinizing all the CAD prompts. This carries the risk that true positive findings will be missed.

The exact impact of CAD on readers’ sensitivity has not been determined. One retrospective study of screening mammograms, involving two different CAD systems, showed that CAD improved the detection of cancers that had been overlooked by a single radiologist. This reduction in false negatives led to an increase in sensitivity from 71.2% for the single radiologist to 84.8% and 80.3% for the two CAD systems.23 Another study, focusing on the detection of previously missed cancers, demonstrated a sensitivity of 51.5% for CAD, 62.5% for the radiologist, and 86.2% for the radiologist and CAD combined.

An emerging issue is that many CAD studies assume, incorrectly, that all cancers prompted by CAD would be interpreted correctly in the clinical environment, leading to an overestimation of CAD’s accuracy. Studies of CAD in routine breast assessment, rather than in the screening scenario, have been advocated to provide a more realistic measure of its impact on the accuracy of mammography in clinical practice.

We evaluated the impact of CAD for full-field digital mammography in a clinical scenario shortly after the introduction of the software at our institution.

 A population of 93 patients (372 digital mammograms) was selected retrospectively. The cases comprised 23 patients with histologically proven malignant lesions (73.9% masses, 26.1% microcalcifications), 31 with benign lesions, and 39 with no lesions. The images were evaluated independently by six blinded radiologists of varying experience, before and after applying the CAD software.

Our results confirmed the general trend discussed above for the screening scenario. Average reader sensitivity for all types of lesions-masses and microcalcifications-varied according to experience. The sensitivity of inexperienced and mildly experienced radiologists to all types of lesions and to masses alone showed a statistically significant increase (p <0.05) when CAD was applied. The use of CAD also increased the false-positive rate slightly (p >0.05).

In conclusion, CAD is still a controversial matter. Most evidence supports its use as a second reader for breast cancer screening. CAD provides adequate sensitivity for the detection of breast cancer, especially microcalcifications, and is particularly helpful when used by less experienced radiologists. The software tools remain less sensitive to masses than to microcalcifications, especially in dense breasts. CAD’s Achilles heel is its related decrease in specificity, and this must be considered when contemplating its use as a screening tool.

Articles

Computer-aided detection (CAD) tools use software to analyze digital or digitized images to find features associated with the target disease.

CAD comes under scrutiny  in breast screening debate

"Perhaps it is for all these reasons, as well as on account of the heterogeneity of Crohn's disease and the unpredictability of the human condition that our quest for the Holy Grail of one 'best measure of disease activity' still remains today, in Korelitz's words of 15 years ago, 'a need unfulfilled.'"

Crohn's disease is an inflammatory disorder of the gastrointestinal tract that is characterized by multiform and, as yet, not well-established pathogenesis.

 It is a chronic condition and has a relapsing course. Initial assessment of inflammatory activity may help determine which patients will benefit from prompt and tailored treatment, and this is consequently of paramount importance.

 Further periodic assessments of inflammatory activity are crucial for monitoring the effects of drugs and for surgical therapy planning.

Many instruments measuring disease-specific activity have been evaluated over the past 30 years. To date, however, no diagnostic technique has emerged as a standard reference tool for the reliable and reproducible quantification of inflammatory activity in Crohn's disease. Gastroenterologists typically make this evaluation on the basis of clinical symptoms, physical findings, laboratory parameters, endoscopy, and imaging tests.

The Crohn's disease activity index (CDAI) is the most widely used clinical scoring tool for evaluations of inflammatory activity specific to Crohn's disease. The CDAI is used in everyday clinical practice and in most therapeutic trials, despite being based partially on subjective criteria that represent patients' perception of their disease rather than objective assessment.

Biochemical markers, including white blood count, erythrocyte sedimentation rate, C-reactive protein, and orosomucoids, have been used as references for activity in imaging studies.

The "scoring system for histologic abnormalities in Crohn's disease mucosal biopsy specimens” has been used to quantify the histological activity of the disease.

 This assessment is not generally recommended as a treatment end point due to the poorly investigated correlation between histological improvement and other assessments of disease activity. Other major drawbacks include the potential for sampling errors and inherent variability in biopsy techniques.

 Histology nonetheless shows a closer and less equivocal correlation with morphologic and postcontrast MRI findings than with clinical aspects.

Barium examination, either by enteroclysis or followthrough technique, has traditionally been the gold standard tool for imaging the small bowel. Invasiveness, exposure to ionizing radiation, and the need for extensive bowel preparation, however, limit the use of this technique for the follow-up of disease.

 Conventional radiography is also unable to evaluate the activity and extension of the inflammatory process through and beyond intestinal mucosa.

Doppler ultrasound has proven useful in distinguishing between active and inactive disease, though no conclusions can be drawn from it about the severity of active disease.

 The role of CT in Crohn's disease has been firmly established, though exposure to ionizing radiation limits its widespread use. A correlation between disease activity and the degree of mural enhancement seen on CT has been demonstrated.

Crohn's disease will last throughout a patient's lifetime. The disease manifests at a relatively young age in approximately 50% of cases. Optimal diagnostic workup requires that patients be monitored frequently. Monitoring techniques should consequently be noninvasive or minimally invasive and, above all, patient-friendly.

Although a perfect tool that meets all these requirements is not presently available, MRI has been advocated as one of the most promising techniques for small bowel imaging. The modality can provide an accurate assessment of disease activity, producing multiplanar images with high soft-tissue contrast. No ionizing radiation is involved, making long-term follow-up feasible, and the discomfort and risks are lower than with other methods.

 Disadvantages include the higher costs associated with MRI and the longer examination time, especially when compared with CT.

Ultrafast MR sequences with breath-hold or respiratory triggering can provide high-quality panoramic images of the gastrointestinal tract and extraintestinal, intra-abdominal structures. The introduction of cine-MR mode has added valuable functional information on the motility of affected loops. This is despite the lower temporal resolution compared with conventional enteroclysis. 

 The impact of MRI on evaluations of Crohn's disease is consequently increasing, and imaging protocols, including small bowel distension techniques (enteroclysis versus followthrough) and MR sequences, have been standardized.

MRI assessment of active disease has been shown to influence patient management significantly, regardless of the number of segments involved,

as therapy is tailored toward individual patients rather than individual segments. MRI features of activity are consequently analyzed in most studies on a per-patient basis.

Visualizing changes in the appearance of the bowel wall helps confirm the presence and phase of disease.

• Bowel wall thickness. Bowel wall thickness has been widely tested as a way of discriminating between active and inactive Crohn's disease (Figure 1).

 One evaluation of therapy response in patients with active Crohn's disease found a substantial decrease of bowel wall thickness. Median values of wall thickness in the small bowel before and after treatment were 4.9 mm and 4.4 mm, respectively (

 Researchers in a separate study concluded that on a per-patient basis, a bowel wall thickness of 4 mm or more has a sensitivity of 88% and a specificity of 75% for active Crohn's disease.

Another group of investigators noted a significant difference in bowel wall thickness (

 < 0.001) within affected bowel segments before and after remission (6.93 ± 3.02 mm versus 4.91 ± 2.31 mm, respectively). However, diseased segments remained pathologically thickened in both phases of disease compared with a control group of individuals with no Crohn's disease who underwent colonoscopy for other reasons (mean wall thickness = 3.37 ± 0.88 mm; p < 0.001). The authors attributed this to the fact that thickening is related to both an acute (edema) and a reversible (inflammation) component of the disease and a chronic component (fibrosis) that does not reverse.

Other investigators have reported a statistically significant and good correlation (

 = 0.0006; k = 0.786) between bowel wall thickness and disease activity.

Bowel wall thickness is usually measured manually on monitor images using the workstation distance-measuring tool. Agreement has not yet been reached on the type of images that should be used for making these measurements. Some reports have measured bowel wall thickness on postcontrast T1-weighted images.

 Others do not mention the sequences used.

 Measurements have also been made on TrueFISP coronal and/or axial images (Figure 2).

Reliable measurement of bowel wall thickness requires adequate luminal distension and a real cross-section of the affected loop. Some authors have suggested averaging multiple bowel wall thickness measurements of the same loop to avoid errors from potentially oblique-angled gauge distances (i.e., measurements not perfectly orthogonal to the real wall thickness).

• Wall contrast enhancement. Enhancement of signal intensity in the bowel wall on T1-weighted images following administration of a gadolinium-based agent has been studied as an indicator of inflammation (Figure 3). The reason for this phenomenon is assumed to be an increase in capillary permeability, blood flow, and local tissue perfusion caused by vasodilatation and neovascularization that leads to an increased interstitial fluid uptake.

 Significantly higher wall contrast enhancement in segments of actively inflamed bowel compared with normal-appearing bowel loops has been reported (

 Another study revealed a significant decrease in the mean value of wall contrast enhancement when patients with Crohn's disease changed from the active phase into clinical remission.

Substantial agreement has been reached that wall contrast enhancement is probably the parameter most closely related to the degree of inflammation. Subjective categorical scales

 have been proposed to reliably quantify this finding. An investigation of postcontrast dynamic measurements, rather than simple enhancement on static images, has shown that both these kinds of MRI analyses can provide only a weak to moderate indication of the severity of Crohn's disease.

• Layered enhancement pattern. A layered pattern of wall contrast enhancement has been observed in bowel segments with proven active disease, leading researchers to suggest that this finding is highly specific for active inflammation. 

 This appearance is thought to result from edema in the submucosal layer (where early inflammatory changes occur in Crohn's disease) and the muscolaris layer, which would lead to more apparent enhancement over the mucosal and serosal surfaces (Figure 3). The observation is supported by another study in which active disease presented as a layered pattern of wall contrast enhancement, and contrast uptake was homogeneous when the disease was inactive. Active disease was detected with high sensitivity (100%), specificity (87.5%), and accuracy (93.8%) on a perpatient basis.

• Further mural findings. The presence of edema in inflamed bowel loops is represented by focal areas with high signal intensity on T2-weighted MRI. A significant correlation between active disease and marked wall hyperintensity has also been found.

 The identification of deep ulcers, a classic finding on conventional enteroclysis, has been shown to be the most sensitive MRI finding for Crohn's disease activity.

Information from beyond the small bowel wall can further refine assessment of Crohn's disease.

• Alteration of fibrofatty tissue. High T2 signal intensity has been reported at the level of fibrofatty tissue in cases of biologically active Crohn's disease.

 This feature is probably related to local mesenteritis, which leads to edema and dilatation of local vessels. An alternative explanation is that the fibrotic component of fibrofatty proliferation is seen as an area of low signal intensity on T2- weighted images, abnormally separating adjacent bowel loops.

Although fibrofatty proliferation is observed frequently in patients with active Crohn's disease, three studies have rated its prevalence as low or having no statistical significance.

• Vascularity of the mesentery. Increased mesenteric vascularity corresponds to the so-called comb sign. T1- weighted images demonstrate short, parallel, linear structures of low signal intensity precontrast or high signal intensity postcontrast close to the mesenteric border and perpendicular to the longitudinal axis of small bowel loops (Figure 4). Two separate studies found that although this feature is more frequently observed in patients with active Crohn's disease, the difference between active and inactive disease did not reach statistical significance.

 One explanation is that the comb sign has two different pathologic substrates in active disease and in Crohn's disease at remission, reflecting an expression of vasa recta dilatation and perivascular fibrotic changes respectively.

• Further perienteric findings. Investigators have observed a homogeneous contrast enhancement (ranging from moderate to intense) in the mesenteric lymph nodes of all patients with a CDAI greater than 150 (active Crohn's disease). This finding was also present in 50% of the patients whose CDAI was less than 150 (inactive Crohn's disease).

 Perienteric changes such as mesenteric lymphadenopathy, fistulas, and abscesses are specific but infrequent among patients with active disease.

Bowel wall thickness and wall contrast enhancement are the features of Crohn's disease on MRI that have been investigated the most and shown to be most reliable. Some authors have attempted to arrange sparse MRI findings into scoring systems capable of providing an overall interpretation.

 Their results suggest that overall assessment of disease activity needs to be based on integrated rather than a few single findings.

Crohn's disease is an inflammatory disorder ofthe gastrointestinal tract that is characterizedby multiform and, as yet, not well-establishedpathogenesis.

Lorenzo Cereser, MD

Articles

CAD comes under scrutiny in breast screening debate

MR imaging features combineto classify Crohn's disease