Blue Cross of Idaho Logo

Express Sign-on

Thank you for registering with Blue Cross of Idaho

If you are an Individual or Family Member under age 65, please register here.

If you are an Medicare or Medicare Supplement member, please register here.

New Options for Affordable Health Insurance

MP 6.01.45

Computer-Aided Evaluation of Malignancy with Magnetic Resonance Imaging of the Breast


Medical Policy    

Section
Radiology

Original Policy Date
10/10/06

Last Review Status/Date
Reviewed with literature search/1:2013

Issue
1:2013

 

Return to Medical Policy Index


Disclaimer

Our medical policies are designed for informational purposes only and are not an authorization, or an explanation of benefits, or a contract.  Receipt of benefits is subject to satisfaction of all terms and conditions of the coverage.  Medical technology is constantly changing, and we reserve the right to review and update our policies periodically.


Description

The use of computer-aided evaluation (CAE) may assist radiologists’ interpretation of contrast-enhanced magnetic resonance imaging (MRI) of the breast, and improve the accuracy of diagnosis of malignancy.

The use of computer-aided evaluation (CAE) is proposed to assist radiologists’ interpretation of contrast-enhanced magnetic resonance imaging (MRI) of the breast. The use of computer-aided evaluation (CAE) is proposed to assist radiologists’ interpretation of contrast-enhanced magnetic resonance imaging (MRI) of the breast. MRI of the breast is suggested as an alternative or adjunct to mammography or other screening and diagnostic tests because of its high sensitivity in detecting breast lesions. However, it has a high false positive rate because of the difficulty in distinguishing between benign and malignant lesions. MRI may be used to screen women at high genetic risk of breast cancer or to look for more extensive disease in women diagnosed with breast cancer who are eligible for breast-conserving surgery; it is also being studied to gauge the impact of cancer treatment (for a discussion of other potential indications, see policy No. 6.01.29, “MRI of the breast”). The CAE systems reviewed in this policy are intended to improve the specificity of MRI in detecting or measuring malignant tissue, while maintaining the generally high sensitivity of MRI. This could potentially reduce biopsy rates if it improves the ability to identify which MRI-detected lesions are almost certainly benign. There is anecdotal information that MRI may also be used in an effort to reduce re-operation rates among patients undergoing breast-conserving surgery by more clearly identifying the tissue that should be removed. The use of CAE may also shorten the time needed to interpret breast MRI images, which currently takes longer than reading mammograms.

CAE systems for MRI essentially provide easier ways of interpreting the patterns of contrast enhancement across a series of images, which in turn may help identify lesions and their likelihood of being malignant. Two key aspects of enhancement (also called kinetics) are examined: 1) within the first minute or so, does the lesion enhance up to a certain threshold (e.g., 50%, 100% of the initial value)? and 2) what is the subsequent pattern of enhancement (continues to increase, plateaus, or declines [called “washout”])? In contrast to computer-aided detection (CAD) systems used with mammography, CAE for MRI is not aimed primarily at identifying lesions for consideration by a radiologist. Unlike the subtle appearance of lesions on mammography, most cancers enhance on MRI. The challenge is determining which lesions are benign and which are malignant. A large number of images are produced during MRI of the breast: images are taken at varying “depths” throughout each breast multiplied by the number of times the breast is imaged to capture different time points in the enhancement process; this can produce hundreds of images. Radiologists view the images to detect suspicious areas, and then they can pick a region of interest and look at the enhancement pattern. However, there may be variations across radiologists in the regions of interest selected and in the precise definition of the region of interest. CAE systems, in contrast, use color-coding and differences in hue to indicate the patterns of enhancement for each pixel in the breast image, thereby allowing the radiologist to analyze the enhancement patterns systematically. CAE systems for MRI of the breast were initially called CAD systems, the same terminology used for mammography. However, the focus with MRI of the breast is on improving specificity (distinguishing malignant from benign) rather than increasing sensitivity (i.e., detection), as in mammography. The authors of 2 recent studies refer to CADStream as a computer-aided evaluation (CAE) program, and that terminology has been adopted in this policy.

Regulatory Status

Several CAE systems for use with MRI of the breast have 510(k) marketing clearance from the U.S. Food and Drug Administration (FDA). The 3TP Software Option, manufactured by 3TP LLC (now called CAD Sciences, White Plains, NY), was cleared on June 23, 2003. iCAD acquired CAD Sciences in 2008 and is now marketing a system called SpectraLook™ with CADVue™. CADstream™, which is manufactured by Confirma, Inc. (Kirkland, WA), was cleared on July 30, 2003; Merger Healthcare (Hartland, WI) subsequently acquired Confirma. A third system called Aegis (Sentinelle Medical Inc, Toronto, Ontario, Canada) received 510(k) marketing clearance from the FDA on February 9, 2007, as substantially equivalent to CADStream Version 4.0. However, in the 510(k) documents, the manufacturer states that the primary goal of Aegis is “to identify where and how deep a biopsy or localization needle should be inserted into an imaged breast.” Additional products include DynaCAD (MRI Devices Corporation, Waukesha, WI; now apparently from Invivo Corp, Orlando, FL), which was cleared July 21, 2004; and Z3D Contrast Acuity Software (Clario Medical Imaging, Inc., Seattle, WA), which was cleared September 5, 2008, and is apparently used in conjunction with CAE for MRI systems. Some of these systems may have broader uses beyond breast MRI. There also may be some overlap in the functions performed by these devices and other image-processing systems.

According to documents filed with the FDA, the 3TP Software Option is “intended to be used as a post-processing software package designed to provide a reliable means for visualizing the presence and pattern of contrast-induced enhancement on MR datasets.” It provides a color-coded image that indicates the likelihood that each pixel shows malignant or benign tissue based on the changes in enhancement at 3 points in time, which are defined by the software program. CADstream is described as a “Computer Aided Detection (CAD) system intended for use in analyzing magnetic resonance imaging (MRI) studies. CADstream automatically registers serial patient image acquisitions to minimize the impact of patient motion, segments and labels tissue types based on enhancement characteristics (parametric image maps), and performs other user-defined post-processing functions (image subtractions, multiplanar reformats, maximum intensity projections). When interpreted by a skilled physician, this device provides information that may be useful in screening and diagnosis…Patient management should not be based solely on the results of the CADstream analysis.” It also provides automated determination of volumes of interest. In addition, CADstream can be used during MRI-guided biopsies.


Policy

The use of computer-aided evaluation (CAE) for interpretation of magnetic resonance imaging (MRI) of the breast is considered investigational.


Policy Guidelines  

 

There is an add-on CPT category III code for the use of computer-aided evaluation (CAE) with MRI of the breast:

0159T Computer-aided detection, including computer algorithm analysis of magnetic resonance imaging (MRI) data for lesion detection/characterization, pharmacokinetic analysis, with further physician review for interpretation, breast MRI.

This code would be used with the code for breast MRI –77058-77059.


Benefit Application
BlueCard /National Account Issues  

 

State or federal mandates (e.g., FEP) may dictate that all FDA-approved devices may not be considered investigational and, thus, these devices may be assessed only on the basis of their medical necessity.


Rationale 

 

This policy regarding computer-aided evaluation (CAE) for magnetic resonance imaging (MRI) of the breast was originally created in 2006 based on a TEC Assessment (1) and has since been updated periodically with a literature search. The most recent update with literature review covers the period from November 2011 through November 2012.

In order to demonstrate the impact of CAE in the diagnosis of breast cancer, studies are needed that compare the sensitivity and specificity of MRI with and without the use of CAE systems. These types of studies can demonstrate the incremental diagnostic accuracy of CAE compared to no CAE. Ideally, these studies should be prospective and should evaluate a population of patients that is similar to that presenting for breast cancer screening in a clinical setting.

In order to demonstrate clinical utility, prospective studies are needed that evaluate whether the incremental diagnostic accuracy leads to changes in management that improve outcomes. Changes in management might include changes in the decision to perform biopsies, which would lead to additional management decisions based on biopsy results.

Literature Review

The literature review focuses on studies that compare the diagnostic accuracy of MRI with and without CAE. There are no prospective studies of this type identified in the literature. The retrospective studies generally do not include a population that is similar to that presenting in clinical care. Rather, they utilize an enriched population that includes a greater proportion of cancers than would be expected in consecutive patients presenting for screening. A representative sample of these studies is discussed below.

What is the incremental improvement in diagnostic accuracy for breast cancer when CAE is added to standard MRI?

The 2006 TEC Assessment summarized 4 published articles and 4 abstracts that compared the accuracy of MRI with and without CAE. (1) (The reviewed studies focused on commercially available CAE systems, but some articles on other systems were included. In addition, studies had to report on cancer detection based on histologic results. Three of the articles reported on development and validation of CAE systems aimed at distinguishing between malignant and benign lesions; and they used information on women with known lesions. The fourth article (2) provided information on one of the non-commercial systems used to evaluate women with cancer who were eligible for breast-conserving therapy (BCT). Additional findings (other lesions or larger lesions) were found in 48 of the 116 (41%) women; approximately 80% of these women had further workup; and in 27 of these women, the findings were malignant. The area under the receiver operating characteristic (ROC) curve was 0.91+/-0.04 for the radiologist reading and 0.98+/-0.04 for the combined radiologist and computerized reading (p=0.03). However, the ability to generalize these results and the clinical impact of the findings is uncertain.

Four abstracts of studies were included in the TEC Assessment because of the small number of studies identified. However, the need to exercise caution in using results from abstracts must be kept in mind as these results are reviewed. Of the 4 abstracts, 2 used CADstream, 1 did not report the system used, and 1 was an excerpt from an article that summarized the results of 3 earlier abstracts on the 3TP system. It is not clear whether the current 3TP system has been modified substantially from the version used in these studies. Once again, these abstracts report on the results of CAE with MRI among women with known lesions.

Finally, DeMartini et al. reported on the use of CAE with MRI in 15 patients to assess the impact of chemotherapy. (3) This small study found there were a substantial number of false negative results for residual malignancy using CAE—a different type of problem than found with most other uses of MRI, i.e., too many false positive results.

Conclusions of the TEC Assessment were that the literature on CAE with MRI of the breast was sparse overall, and that few studies addressed the specific situations in which CAE with MRI is used in a clinical setting. The few articles and abstracts calculated test characteristics on the basis of lesions and not the number of women or breasts. In a screening population, many women would not have any lesions. Including these women might alter the results. Given MRI’s lower sensitivity in detecting ductal carcinoma in situ (DCIS), the mix of DCIS versus masses would affect the calculations of sensitivity and specificity and might affect the impact of the CAE system.

Since the TEC Assessment was completed in 2006, there have been a number of relevant studies published. All of these have been retrospective analyses that included populations of patients that are not representative of those seen in clinical care.

Two articles were published that were apparently based on one of the retrospective studies presented in an abstract form from the TEC Assessment. The first article, published in 2006, reported on 33 consecutive lesions biopsied under MRI guidance at a single institution. (4) The second article, published in 2007, reported on 155 consecutive lesions that appeared to subsume the 33 lesions included in the 2006 study; the later article is therefore summarized here. (5) The lesions were not palpable or visible on mammography or sonography and were assessed with and without CAE. (4) All of these lesions were rated Breast Imaging Reporting and Data Systems (BIRADS) 4 or 5, i.e., suspicious or highly suggestive of malignancy. Of these lesions, 64% were in recently diagnosed breast cancer patients, 14% were in high-risk patients being screened, and 14% were for problem solving. Three different MRI imaging protocols were used. CADStream was then retrospectively applied for this study. As expected, increasing the level of enhancement required (to 100%) lowered the number of false positive results. At the 50% enhancement level, no statistically significant difference was found in the positive predictive value between the initial reading and the subsequent application of CAE. At the 100% enhancement level, however, the positive predictive value was significantly higher with CAE than without (30.4% vs. 26.6%, respectively, p=0.02). Because the radiologists who read each set of images with and without CAE were not necessarily the same, it is possible that some of this difference might be due to a variation across readers rather than to the addition of CAE. There was no significant difference in subsequent enhancement patterns (i.e., washout, persistent, or plateau) between benign and malignant lesions; and many lesions included diverse enhancement patterns.

Another retrospective study evaluated the use of the CADStream system, using images from a 3.0 T MRI system. (6) In this retrospective analysis of 426 women imaged consecutively, only 36 patients with 42 lesions were used in the final analysis. The final sample included women with indeterminate mammographic and/or ultrasound results or high-risk screening. Images were interpreted by two experienced breast radiologists and two residents. Manual reading was followed 6 months later with reading of the CAE results. The sensitivity and specificity of the manual reading for both experienced radiologists combined was 84.6% and 68.8%, respectively (residents only read CAE results; BIRADS rating of 3 considered negative). The sensitivity and specificity of the CAE reading for the same two experienced radiologists was 90.4% and 81.3%, respectively; the difference in specificity manually versus with CAE was statistically significant at p<0.05. There were no statistically significant differences among the CAE readers, including residents versus experienced breast radiologists. While the results are interesting, this study has several limitations, including its retrospective nature, highly selective sample with a large proportion of cancer cases, and small number of readers. It also is not clear whether the results apply only to a 3.0 T MRI system. Further research is needed with larger, more robust studies to assess these findings.

Arazi-Kleinman et al. performed a retrospective study that evaluated the sensitivity and specificity of a new MRI CAD software prototype called CAD-Gaea (Sentinelle Medical, Toronto, Canada; commercially released under the name Sentinelle Aegis software). (7) The patients were women at high risk of breast cancer (BRCA1, BRCA2, or calculated lifetime risk of being a mutation carrier of >25%). From an initial sample of 1,548 MRI studies on a 1.5 T system, the study sample consisted of 56 lesions in 53 women. Thirty-nine percent of the lesions were malignant, and of those, 59% were DCIS. The utility of this study for the present policy is limited because it only includes BIRADS 3-5 cases that were biopsied (as well as a fat suppression protocol). Readings were performed by two experienced breast radiologists looking at different aspects of the CAE results, e.g., different thresholds for initial enhancement and initial enhanced versus delayed pattern or both. The primary finding was that high levels of sensitivity could be achieved using CAE (both initial and delayed enhancement patterns) for invasive cancers (100%), but when DCIS was included, the sensitivity dropped to 73%. The specificity when including all lesions was 56%. The prospective radiologist interpretation (apparently as part of clinical care) was more sensitive than the CAE-based interpretation (p=0.05) but less specific (p=0.01); specific values were not reported. The authors conclude that “The breast MRI CAD system used could not improve the radiologists’ accuracy for distinguishing all malignant from benign lesions, due to the poor sensitivity for DCIS detection.”

Shimauchi et al. (8) performed a retrospective analysis in 2011 that was similar to previous studies. In this study, an initial training set of 121 breast lesions (77 malignant and 44 benign) was used to train the CAD system. A test sample of 30 malignant breast lesions and 30 benign lesions was used to test the incremental accuracy of computer-aided detection (CAD) when added to conventional MRI imaging. The sensitivity for detection of malignancy was higher for CAD compared to conventional imaging (88% vs. 83%, p=0.001), as was the mean area under the curve by ROC analysis (84% vs. 80%, p=0.007). The mean specificity was not significantly greater in the CAD group compared to conventional imaging (53% vs. 50%, p=0.20).

Cho et al. performed a retrospective study in 2012 that evaluated CAE in detection of contralateral breast cancer lesions in women with breast cancer. (9) There were a total of 23 malignancies and 29 benign lesions in 52 consecutive patients. Three experienced radiologists interpreted the images and provided their judgments about the probability of cancer. Sensitivity and specificity was determined with and without CAE for each radiologist. The mean sensitivity was significantly improved with CAE, from 76.8% to 92.8%, but the specificity was decreased by a non-significant amount (40.2% to 35.6%). There was an increase in the combined accuracy, as measured by the area under the curve from 0.603 to 0.667, but this difference was not statistically significant.

Research continues on efforts to improve the ability of CAE systems to provide information that increases diagnostic accuracy. These include articles that focus on the use of different metrics (e.g., selecting the “most suspect” enhancement pattern of a lesion versus the distribution of different enhancement patterns within a lesion (10-12); use of textural analysis of lesions (13); and expansion of the use of CAE to help distinguish the type of breast cancer and whether it has metastasized (14)). These studies did not necessarily use commercially available CAE systems. Further research is needed to evaluate the utility of all of these approaches.

What is the clinical utility of CAE when added to standard MRI of the breast?

There is no direct evidence that evaluates the impact of CAE on health outcomes. There are no prospective studies that examine whether management decisions are changed due to results of CAE. There are also no relevant modeling studies that estimate the impact of CAE on outcomes.

Decisions for biopsies may be changed as a result of CAE; in particular, biopsies may be performed in areas of abnormality identified by CAE that were not seen on standard MRI. This may in turn improve the detection rate for malignancies. It is also possible that the number of false positive biopsies is increased when CAE is used. Since the incremental changes in sensitivity and specificity with CAE are not known, it is not possible to estimate the number of additional malignancies that would be detected by CAE, nor is it possible to determine the number of additional false positive biopsies that would be performed. As a result, the clinical utility of CAE when added to standard MRI of the breast has yet to be determined.

Summary

The available evidence consists primarily of retrospective studies that compare the accuracy of computer-aided MRI imaging of breast malignancy versus conventional imaging. The populations in these studies are not representative of patients seen in clinical care, rather they include samples of women who are highly selected and usually have far more cases of cancer than would be encountered in a screening population. As a result, the true sensitivity and specificity of computer-aided MRI imaging, and the incremental improvement in accuracy over conventional imaging, cannot be determined with certainty. Larger, well-designed, prospective studies are needed that include relevant clinical populations in order to determine whether computer-aided evaluation results in a clinically significant improvement in diagnostic accuracy. As a result of the deficiencies in the available literature, the use of computer-aided evaluation of malignancy with MRI imaging is considered investigational.

Practice Guidelines and Position Statements

The European Society of Breast Cancer Specialists (EUSOMA) issued consensus recommendations for MRI imaging of the breast in 2010. (15) This document states “We recommend the use of standardized interpretation systems such as the above mentioned BI-RADS lexicon, or equivalent. There is some evidence that software for breast MR computer-aided diagnosis (CAD) may be of benefit but insufficient to recommend the routine use of such systems.”

Medicare National Coverage

None

References:

    1. Blue Cross and Blue Shield Association Technology Evaluation Center (TEC). Computer-Aided Detection of Malignancy with Magnetic Resonance Imaging of the Breast. TEC Assessments 2006; Volume 21, Tab 4.
    2. Deurloo EE, Peterse JL, Rutgers EJ et al. Additional breast lesions in patients eligible for breast-conserving therapy by MRI: impact on preoperative management and potential benefit of computerised analysis. Eur J Cancer 2005; 41(10):1393-401.
    3. Demartini WB, Lehman CD, Peacock S et al. Computer-aided detection applied to breast MRI: assessment of CAD-generated enhancement and tumor sizes in breast cancers before and after neoadjuvant chemotherapy. Acad Radiol 2005; 12(7):806-14.
    4. Lehman CD, Peacock S, DeMartini WB et al. A new automated software system to evaluate breast MR examinations: improved specificity without decreased sensitivity. AJR Am J Roentgenol 2006; 187(1):51-6.
    5. Williams TC, DeMartini WB, Partridge SC et al. Breast MR imaging: computer-aided evaluation program for discriminating benign from malignant lesions. Radiology 2007; 244(1):94-103.
    6. Meeuwis C, van de Ven SM, Stapper G et al. Computer-aided detection (CAD) for breast MRI: evaluation of efficacy at 3.0 T. Eur Radiol 2010; 20(3):522-8.
    7. Arazi-Kleinman T, Causer PA, Jong RA et al. Can breast MRI computer-aided detection (CAD) improve radiologist accuracy for lesions detected at MRI screening and recommended for biopsy in a high-risk population? Clin Radiol 2009; 64(12):1166-74.
    8. Shimauchi A, Giger ML, Bhooshan N et al. Evaluation of clinical breast MR imaging performed with prototype computer-aided diagnosis breast MR imaging workstation: reader study. Radiology 2011; 258(3):696-704.
    9. Cho N, Kim SM, Park JS et al. Contralateral lesions detected by preoperative MRI in patients with recently diagnosed breast cancer: application of MR CAD in differentiation of benign and malignant lesions. Eur J Radiol 2012; 81(7):1520-6.
    10. Baltzer PA, Renz DM, Kullnig PE et al. Application of computer-aided diagnosis (CAD) in MR-mammography (MRM): do we really need whole lesion time curve distribution analysis? Acad Radiol 2009; 16(4):435-42.
    11. Yuan Y, Giger ML, Li H et al. Multimodality computer-aided breast cancer diagnosis with FFDM and DCE-MRI. Acad Radiol 2010; 17(9):1158-67.
    12. Ko ES, Choi HY, Lee BH et al. Central washout sign in computer-aided evaluation of breast MRI: preliminary results. Acta Radiol 2011; 52(3):256-63.
    13. Holli K, Laaperi AL, Harrison L et al. Characterization of breast cancer types by texture analysis of magnetic resonance images. Acad Radiol 2010; 17(2):135-41.
    14. Bhooshan N, Giger ML, Jansen SA et al. Cancerous breast lesions on dynamic contrast-enhanced MR images: computerized characterization for image-based prognostic markers. Radiology 2010; 254(3):680-90.
    15. Sardanelli F, Boetes C, Borisch B et al. Magnetic resonance imaging of the breast: recommendations from the EUSOMA working group. Eur J Cancer 2010; 46(8):1296-316.
      

Codes

Number

Description

CPT 

0159T 

Computer-aided detection, including computer algorithm analysis of MRI image data for lesion detection/characterization, pharmacokinetic analysis, with further physician review for interpretation, breast MRI  

ICD-9 Procedure 

 

 

ICD-9 Diagnosis 

174.0 – 174.9 

Malignant neoplasm of female breast 

 

175.0 – 175.9 

Malignant neoplasm of male breast 

 

198.81 

Secondary malignant neoplasm of breast 

 

233.0 

Carcinoma in situ of breast 

 

611.72 

Lump or mass of breast 

 

V10.3 

Personal history of breast cancer 

 

V16.3 

Family history of breast cancer 

HCPCS 

 

 

ICD-10-CM (effective 10/1/13)   Investigational for all diagnoses
  C50.01-C50.929 Malignant neoplasm of breast code range
  V79.81 Secondary malignant neoplasm of breast
  D05.01-D05.99 Carcinoma in situ of breast
  N63 Unspecified lump in breast
  Z80.3 Family history of malignant neoplasm of breast
  Z85.3 Personal history of malignant neoplasm of breast
ICD-10-PCS (effective 10/1/13)   No code specific to CAD
  BH30Y0Z, BH30YZZ, BH30ZZZ, BH31Y0Z, BH31YZZ, BH31ZZZ, BH32Y0Z, BH32YZZ, BH32ZZZ Magnetic resonance imaging, breast(s), code list

Type of Service 

Radiology 

Place of Service 

Outpatient 


Index

CAD, Magnetic Resonance Imaging (MRI) Breast
Computer-Aided Detection with Magnetic Resonance Imaging (MRI), Breast
Magnetic Resonance Imaging (MRI), Breast, with Computer-Aided Detection


Policy History

 

Date

Action

Reason

10/10/06

Add to Radiology section

New policy

04/09/08 Replace policy   Policy updated. Title revised- “detection” changed to “evaluation”. The acronym “CAD” changed to “CAE” throughout policy. Reference includes more information, references 4-5 added. Policy statement unchanged.  
04/24/09 Replace policy  Policy updated with search with literature search in March 2009; policy statement unchanged 
06/10/10 Replace policy Policy updated with literature search; references 6-10 added. Policy statement unchanged
1/12/12 Replace policy Policy updated with literature search; references 8,10,11,14 added. Policy statement unchanged
1/10/13 Replace policy Policy updated with literature review, reference 9 added. Policy statement unchanged.