Clinical Testing of High-Spatial-Resolution Parametric Contrast-Enhanced MR Imaging of the Breast

HOME

HELP

FEEDBACK

SUBSCRIPTIONS

Clinical Testing of High-Spatial-Resolution Parametric Contrast-Enhanced MR Imaging of the Breast

Frederick Kelcz¹, Edna Furman-Haran², Dov Grobgeld² and Hadassa Degani²

^¹ Department of Radiology, University of Wisconsin Hospital and Clinics, E3/311-3252 Clinical Sciences Center, 600 Highland Ave., Madison, WI 53792-3252.
^² Department of Biological Regulation, Weizmann Institute of Science, P. O. Box 26, Rehovot, Israel 76100.

View larger version (10K):

[in a new window]

Fig. 1A. —Illustration of how time points are selected for three-time-point method. Modeling parameter K refers to the vascular permeability x surface area product per unit volume. Modeling parameter EVF refers to extracellular volume fraction. Graph shows hypothetic signal intensity (SI) enhancement versus time profiles for malignant (red curve) and benign (blue curve) lesions after gadopentetate dimeglumine injection. In this simulation, values used for malignant lesion were K = 0.95 min^-1, EVF = 0.5; and for benign lesion, K = 0.3 min^-1, EVF = 0.5. Curves were generated using model-based equation that accounted for MR imaging sequence parameters and T1 relaxation rate, as well as dose, relaxivity, and blood pharmacokinetics of contrast agent. Vertical dotted lines simulate SI measurements taken at selected time points. Note rules by which three-time-point method codes spatial coordinate as red, green, or blue on basis of SI differences between second and third time points: red, SI decrease; green, SI stable; blue, SI increase.

View larger version (32K):

[in a new window]

Fig. 1B. —Illustration of how time points are selected for three-time-point method. Modeling parameter K refers to the vascular permeability x surface area product per unit volume. Modeling parameter EVF refers to extracellular volume fraction. Calibration map results show SI measurements at 0, 0.5, and 6 min. Simulated benign and malignant lesion values of K and EVF referred to in A are shown as points B and M, respectively. Three-time-point calibration map shows assignment of overlay color based on SI measurement changes. Choice of inappropriately early second time point (0.5 min) leads to both benign and malignant simulated lesions being coded blue, or benign, because SI of malignant lesion has not yet peaked (0.5 min < its SI at 6 min). Superimposed yellow curves are isoslopes of wash-in values—the faster the wash-in, the brighter the overlay color.

View larger version (29K):

[in a new window]

Fig. 1C. —Illustration of how time points are selected for three-time-point method. Modeling parameter K refers to the vascular permeability x surface area product per unit volume. Modeling parameter EVF refers to extracellular volume fraction. Calibration map results show SI measurement at 0, 2, and 12 min. Choice of inappropriately late third time point (12 min) leads to benign lesion being coded as indeterminate—in green rather than in blue zone (confident benign). Malignant lesion is appropriately coded as red.

View larger version (30K):

[in a new window]

Fig. 1D. —Illustration of how time points are selected for three-time-point method. Modeling parameter K refers to the vascular permeability x surface area product per unit volume. Modeling parameter EVF refers to extracellular volume fraction. Calibration map results show SI measurements at 0, 2, and 6 min. Optimal choice of both second and third time points (2 and 6 min, respectively) produces desired result—correct discrimination and color coding of malignant lesion as red and benign lesion as blue. Note correct choice of time points typically equally divides K-EVF plane into symmetric zones of equal red and blue. Width of green (indeterminate) area is governed by definition of lack of defined change in slope between second and third time points (in our case, ± 10%). Once this calibration map was set for MR imaging parameters used in our study, data for all clinical patients were taken at same time points and used same rules to reach diagnosis.

View larger version (105K):

[in a new window]

Fig. 2A. —69-year-old woman with island of benign breast tissue simulating mass (true-negative study). Mediolateral oblique mammogram shows ovoid focal abnormality (arrow) along inferior mid breast.

View larger version (87K):

[in a new window]

Fig. 2B. —69-year-old woman with island of benign breast tissue simulating mass (true-negative study). Three-time-point sagittal parametric MR image shows virtually all dark blue pixels (more apparent on computer screen), indicating lesion with low values of K (vascular permeability x surface area product per unit volume) and EVF (extracellular volume fraction), which are indicators of benignity. Prospective three-time-point diagnosis was benign (suspicion level 2).

View larger version (137K):

[in a new window]

Fig. 3A. —64-year-old woman with infiltrating ductal cancer (true-positive study). Mediolaterial oblique mammogram shows area of ill-defined density and architectural distortion (arrow).

View larger version (117K):

[in a new window]

Fig. 3B. —64-year-old woman with infiltrating ductal cancer (true-positive study). Subtraction MR image (image at 0 min [before administration of gadolinium contrast agent] subtracted from image at 6 min) shows irregular area of enhancement corresponding to mammographic lesion.

View larger version (104K):

[in a new window]

Fig. 3C. —64-year-old woman with infiltrating ductal cancer (true-positive study). Top three are magnified images from adjacent slices of central portion of three-time-point parametric MR image, calculated using 0-, 2-, and 6-min MR images. Results show visually indeterminate number of red pixels. Bottom three images are these same image locations, but three-time-point images were recalculated using 0-, 4-, and 8-min MR images. Note shift toward increasing number of red pixels, indicating malignancy to be more probable than benignity. Prospective three-time-point diagnosis was malignant (suspicion level 4).

View larger version (11K):

[in a new window]

Fig. 4. —Graph of receiver operating characteristic curve derived from three-time-point data in which lesion location was supplied to researcher, but no other clinical information was provided. Results show good discrimination between benign and malignant diagnoses (area under the curve = 0.91).

View larger version (159K):

[in a new window]

Fig. 5A. —60-year-old woman with low- to intermediate-grade ductal cancer in situ (false-negative study). Optical close-up of magnification craniocaudal mammogram shows cluster of microcalcifications in superior mid breast.

View larger version (147K):

[in a new window]

Fig. 5B. —60-year-old woman with low- to intermediate-grade ductal cancer in situ (false-negative study). Optical close-up shows subtraction sagittal MR image 6 min after injection of gadodiamide. Small irregular focus of enhancement corresponds to cluster of microcalcifications.

View larger version (137K):

[in a new window]

Fig. 5C. —60-year-old woman with low- to intermediate-grade ductal cancer in situ (false-negative study). Optical close-up of sagittal three-time-point MR image shows area of predominantly bright green pixels thought to be benign (score, 2). Three of eight ductal carcinoma in situ (DCIS) lesions were misdiagnosed as benign by three-time-point method. Area of future investigation is to determine whether a specific pattern for DCIS will allow increased accuracy of diagnosis.

View larger version (181K):

[in a new window]

Fig. 6A. —45-year-old woman with fibroadenoma (true-negative study). Sonogram of mammographically occult palpable lesion, showing gently lobulated 1.8-cm mass, without acoustic shadowing, typical of fibroadenoma.

View larger version (142K):

[in a new window]

Fig. 6B. —45-year-old woman with fibroadenoma (true-negative study). Subtraction maximum-intensity-projection MR image obtained, 6 min after gadodiamide injection shows that fibroadenoma (arrow) confirmed on sonography is largest of multiple enhancing smaller lesions. At workstation, many of these smaller lesions showed enhancement profile similar to that of larger, palpable, and sonographically confirmed lesion. In clinical practice, evaluation of multiple other enhancing lesions by manual placement of region of interest is impractical. Although internal septations are said to be important MR imaging sign of fibroadenoma, they were not noted in this patient.

View larger version (106K):

[in a new window]

Fig. 6C. —45-year-old woman with fibroadenoma (true-negative study). Three-time-point parametric MR image shows predominantly central bright green and peripheral blue pixels (arrow), consistent with benign lesion (score, 2). Three-time-point parametric image showed no other suspicious lesions; however, because of multiplicity of similar lesions, patient is being followed up. Confirmation of multiple benign lesions will be assumed if no malignancy is diagnosed after 2 years of mammographic follow-up.

View larger version (137K):

[in a new window]

Fig. 7A. —44-year-old woman with multifocal infiltrating ductal cancer (true-positive study). Optical close-up of mammogram (mediolateral oblique projection) shows spiculated mass (arrow). As result of this patient's volunteering for three-time-point clinical trial, second unsuspected mammographically occult lesion was detected at location estimated by arrowhead.

View larger version (71K):

[in a new window]

Fig. 7B. —44-year-old woman with multifocal infiltrating ductal cancer (true-positive study). Three-time-point parametric sagittal plane MR slice shows lesion (arrow) suspected to be malignant. High predominance of bright red pixels indicates high value of K (vascular permeability x surface area product per unit volume) and low EVF (extracellular volume fraction), indicating high probability of malignancy (score, 5).

View larger version (76K):

[in a new window]

Fig. 7C. —44-year-old woman with multifocal infiltrating ductal cancer (true-positive study). Three-time-point parametric MR image of second adjacent sagittal slice shows second site (arrowhead) very suspicious for malignancy (score, 5). Radiologist discussed scan with surgeon, and both sites were biopsied at time of surgery, confirming unsuspected multifocal malignancy. Six of 58 women volunteers are known to have benefited in some fashion from participating in three-time-point clinical trials (one additional patient may have benefited).

CiteULike

Complore

Connotea

Del.icio.us Add to Digg

Digg

Technorati What's this?