Breast imaging. State-of-the-art.

The underlying goal of all breast imaging procedures is the detection of cancer. X-ray mammography places greatest emphasis on the disclosure of early, nonpalpable, curable cancer. Diagnostic ultrasonography stresses the differentiation of benign cysts from diagnostically indeterminate solid masses that require biopsy. Evolving experimental procedures such as transillumination light-scanning and magnetic resonance imaging currently are undergoing preliminary evaluation. An older procedure, thermography, seems to operate at a high level of effectiveness only for advanced cancer. Since x-ray mammography is the technique that has proven most successful in detecting early breast cancer, it is the standard to which all imaging alternatives must be compared.

Digital mammography. A comparison of two digitization methods.

A preliminary study of 40 different radiodense breasts digitized with a Fuji high resolution BAFBr:EU2+ imaging plate enabled us to establish acceptable enhancement procedures with a Fuji Computer Radiology 201 system. Screen-film images of 36 of these breasts were also digitized and enhanced on a Damon DETECT TV system. Three radiologists specializing in mammography reviewed each pair of images. For the 20 normal examinations, both digital methods were considered equivalent in image quality, while for the 16 cases containing pathology (masses and/or calcifications) the TV system was considered to provide the best image quality twice as often as the laser scanned system. The radiologists rejected both methods of enhancement for 8% of the images. Despite cost differences between the two systems, both have equal capability in penetrating dense breasts. However, both systems have several significant deficiencies which preclude their clinical use. At the present time, there is no objective justification for using either system for breast imaging other than in an experimental capacity.

Use of a Doppler guide wire for intravascular blood flow measurements: a validation study for potential neurologic endovascular applications.

PURPOSE: To conduct a validation study of a Doppler guide wire for potential neuroendovascular applications. METHODS: A 12-MHz, 0.018-inch Doppler guide wire was evaluated in eight swine under various blood flow conditions using two types of in vivo cerebrovascular models (physiologic and arteriovenous shunting). Flow conditions were mechanically and pharmacologically altered. Doppler average peak velocity was compared with volumetric blood flow, and flow profile corrections were calculated and analyzed. Qualitative aspects of the Doppler guide wire spectra were also assessed. RESULTS: Plots of average peak velocity versus volumetric blood flow showed excellent linear relationships (r2 > 0.94), which were maintained at high flow conditions (average peak velocity, 99 to 236 cm/sec; volumetric blood flow, 392 to 889 mL/min). Values of flow profile correction varied from 0.43 to 0.94 and showed no consistent relationship to changes in volumetric blood flow. CONCLUSIONS: The excellent correlation between average peak velocity and volumetric blood flow over a wide range of blood flow conditions and the additional qualitative information of the Doppler guide wire spectra establish a foundation for clinical implementation. The unpredictable variations of flow profile corrections remain obstacles for calculating volumetric blood flow based on Doppler guide wire average peak velocity.

Acceptance testing prone stereotactic breast biopsy units.

When the Mammography Quality Standards Act becomes law in October, 1994, stereotactic breast biopsy units may require yearly physicist calibration. Upright stereotactic units can be easily tested using conventional mammography procedures and a gelatin phantom containing simulated calcifications, but prone units are difficult to assess because of the under-table tube configuration. The two current manufacturers of these units have made different design decisions which affect each unit's calibration. There are a number of important distinctions between screening and prone biopsy units. For the two currently available prone units, a pronounced heel effect makes ion chamber position critical. Focal spot measurements are particularly difficult on one unit because there is no light field. The fixed grid on the other unit must be tested with a flood film. Physicists who inspect these units before their clinical use should be aware of variations needed by this equipment for specific acceptance tests.

Mammography focal spot measurement with a star pattern: techniques to avoid inaccuracies.

The increased emphasis on mammography acceptance testing and quality control make the star pattern test an essential component of the procedure for diagnosing poor resolution due to large focal spots. In this paper procedures for calculating magnification and the correct angle of the star pattern to use are described. The preferred placement of the star pattern in the field and proper exposure technique is also discussed.

Focal spot size measurements with pinhole and slit for microfocus mammography units.

Many mammography units now have microfocal spots ranging in size from 0.09 to 0.4 mm for magnification radiography. On site measurements of these focal spots are not possible without invasive procedures if National Equipment Manufactures Association (NEMA) specifications are to be followed. This paper describes a method to perform such measurements expeditiously (1.5 h) and noninvasively. We describe how this method differs from NEMA test specifications, and report the test results of 22 units from ten manufacturers. Six of the focal spots were larger than the manufacturer's specifications which were based on current NEMA standards. Emphasis is placed on the need for verifiable standards for mammography x-ray systems.

Medical pattern recognition pitfalls in a clinical setting: renal cell carcinoma survival prediction.

Three studies to predict renal cell carcinoma patient survival from tumor textures found in digitized early venous phase arteriograms were successful individually. However, when the three methods were compared, they were not consistent and no single method was clinically useful. The first study predicted 5 yr survival of 37 patients with 87% accuracy. The second study added 29 patients to the data base; the poor survival of the 21 patients who died within 5 yr of diagnosis was predicted with 80.9% accuracy. When 27 of these cases were redigitized with a laser scanner, average survival prediction accuracy was 78%. In these studies, digitization hardware, radiographic technique, normalization methods, window selection, and contrast medium distribution all contributed to differences in the statistics separating poor from good patient survival.

Mammography screen-film selection: individual facility testing technique.

Variations in tube output, film processing, and radiologist's preferences affect the screen-film combination that is appropriate for any particular mammographic facility. A technique to test a variety of screen-film combinations for screening mammography is described. Films are selected for testing because of their densitometric characteristics. Dose and clinical reliability are established with phantoms before the screen-film combinations are used to image consecutive patients having bilateral examinations. The mammograms selected for evaluation are those with similar optical density ranges, and which also may be compared to available previous mammograms or which have unusual mammographic findings. All radiologists reading mammograms at a facility independently score the selected cases. Scores of "unacceptable," "acceptable," or "outstanding" are assigned to four basic imaging characteristics: sharpness, contrast, visibility of skin line, and noise. Interobserver variations by this method require normalization, unlike ROC analysis which is not applicable for this data because of the absence of proved pathologic diagnoses. The testing of 5 films and two screens using 42 patient examinations required 2 h of time from each radiologist. It took 7 h of the physicist's time to pretest the 5 films, select the 42 acceptable examinations for testing by the radiologists, and summarize the data.

Film-screen mammography x-ray tube anodes: molybdenum versus tungsten.

Gridless screen-film mammography at 23 kVp with a W anode (inherent filtration: 0.1 mm A1, added filtration: 0.025 mm Mo) can achieve contrast identical to that achieved with gridless film-screen mammography at 27 kVp with a Mo anode (inherent filtration: 1.0 mm Be, added filtration: 0.025-mm Mo). However, W-anode film-screen mammograms obtained at 23 kVp require more radiation than Mo-anode film-screen mammograms obtained at 27 kVp. The lack of contrast of W-anode film-screen images produced at the same kVp as Mo-anode images was verified clinically and with a low contrast test object imaged over a range of densities. A step wedge test object was then used to match contrast between Mo- and W-anode gridless film-screen images at various kVp. The low contrast test object images verified the contrast equivalence of images obtained at 23 kVp for a W anode and 27 kVp for a Mo anode. A comparison of the two kVp for different anode materials was tested clinically on ten patients. The clinical and low contrast test object experiments were reviewed by three radiologists specializing in mammography.

A review of mammography test objects for the calibration of resolution, contrast, and exposure.

Mandated and voluntary accreditation and quality control programs for mammography require the use of standardized mammography test objects. We evaluated eleven commercially available test objects and three prototype test objects, comparing them with respect to their resolution targets, contrast targets, and the dose they required when imaged by the same automatic exposure meter. Ion chamber and/or thermoluminescent dosimeter measurements of exposure were made with each test object, while attenuation was measured for seven. Measurements of dosage using acrylic (5 test objects) and tissue equivalent epoxy (9 test objects) showed as much as a 400% variation in the radiation supplied by the same automatic exposure device when differences in thicknesses of test objects were normalized. Speck visibility was as dependent on the composition of the specks and of the surrounding material as on the size of the specks. Contrast targets were adequate in only three test objects. Optical density differences between images of a 4-cm-thick breast and of different test object materials, also 4 cm in thickness, exposed to the same radiation, imply that untested acrylic or epoxy resin materials should not used in the calibration of automatic exposure controls.

Using tissue texture surrounding calcification clusters to predict benign vs malignant outcomes.

The positive predictive value of mammography is between 20% and 25% for clustered microcalcifications. For very early cancers there is often a lack of concordance between mammographic signs and pathology. This study examines the usefulness of computer texture analysis to improve the accuracy of malignant diagnosis. Texture analysis of the breast tissue surrounding microcalcifications on digitally acquired images during stereotactic biopsy is used in this study to predict malignant vs benign outcomes. 54 biopsy proven cases (36 benign, 18 malignant) are used. The texture analysis calculates statistical features from gray level co-occurrence matrices and fractal geometry for equal probability and linear quantizations of the image data. Discriminant models are generated using linear discriminant analysis and logistic discriminant analysis. Results do not differ significantly by method of quantization or discriminant analysis. Jackknife results misclassify 2 of 18 malignant cases (sensitivity 89%) and 6 of 36 benign cases (specificity 83%) for logistic discriminant analysis. From this preliminary study, texture analysis appears to show significant discriminatory power between benign and malignant tissue, which may be useful in resolving problems of discordance between pathological and mammographic findings, and may ultimately reduce the number of benign biopsies.

Mammography fixed grid versus reciprocating grid: evaluation using cadaveric breasts as test objects.

In this study we use unfixed cadaveric breasts to obtain mammography images with fixed and reciprocating grids. Sheets of acrylic, containing one or more clusters of simulated calcifications and masses, were superimposed on two fresh cadaveric breasts (3.4 and 6.5 cm thick), and were imaged with a fixed grid and a reciprocating grid. Six radiologists, working independently, attempted to identify the number of clusters and/or masses in 114 images containing 139 clusters of simulated calcifications and 42 simulated masses. Thirty-four of these images were normal, containing no lesions. For the thinner breast, no statistically significant difference was found in the detection of clusters of calcifications in the images produced with the fixed grid compared to those produced with the reciprocating grid. However, for the detection of calcifications in images of the thicker breast, sensitivity of 74% for detection of calcifications when a fixed grid was used was significantly less than sensitivity of 86% when a reciprocating grid was used (P = 0.006). The mass detection sensitivity was 91% for images made with a fixed grid compared to 96% for images made with a reciprocating grid, but the difference was not statistically significant (P = 0.346). The use of cadaveric breasts as test objects was well accepted by radiologists. Only for the thick cadaveric breast were differences between the two grids significant, and these differences were restricted to the task of finding calcifications.

Mammography film processor replenishment rate: bromide level monitoring.

The effects of the mammography film processing replenishment rate on contrast and speed are studied sensitometrically. Two experiments studied decreasing replenishment rates in the Kodak RP developer and quantified changes in the developer by measuring bromide ion concentrations. First, values of NaBr concentration from 1.7 to 8.4 g/L, achieved by reducing the replenishment rate, were tested with sensitometry strips. Second, the developer replenishment rate of a high volume dedicated mammography processor was reduced by one-third, to 20 cm3/1560 cm2, so that the NaBr concentration rose from 2.0 to 12.36. Sensitometric results for four film types and patient films were tested for changes from standard values as NaBr concentration was restored to 3.31 g/L. Fifty-five clinical images obtained at 7.3-9.3 NaBr g/L were compared to their matching previous films, with NaBr levels of 2-3 g/L, for contrast and visibility of the skin line. For the range of the NaBr ion from 1.7 to 7 g/L, no significant sensitometric differences were found. Above 7 g/L, different film types had different sensitometric results. From 7.3 to 9.3 NaBr g/L, 47.5% of the clinical films reviewed by four radiologists had less contrast compared to previous films. Dedicated mammography processors with high film volume (i.e., those that do not have excessive oxidation or foreign dye problems) can operate at lower replenishment rates than are currently employed. All common mammography film types are stable at these lower replenishment rates up to 7.0 NaBr g/L.

Diagnostic effects of edge sharpening filtration and magnification on digitally subtracted renal images.

The improved appearance of digital radiographs filtered to improve local contrast and sharpen edges has not increased acceptance of these images by radiologists. Furthermore, many radiologists assert that correct diagnosis is not improved with these filtered images. This study was designed to test this assertion for digital subtraction angiograms (DSA) of renal images. Four experiments are described. First, phantom studies identified filters and their parameters thought likely to be acceptable and useful in diagnosing renal images formed by DSA. Second, these filters and parameters were then tested on medical images to assess their acceptance by radiologists. Third, display modes of windowing, positive/negative presentation, and magnification were varied for filtered and unfiltered images to assess preferences of radiologists. Fourth, filtered and unfiltered magnified images were used to test improved diagnosis. In the final experiment, 148 images from 33 renal studies (15 normal, 18 abnormal) were magnified, gray level windowed, and filtered. Diagnosis was not improved by the two edge sharpening filters tested.

Testing mammography equipment: evolution over a 4-year period.

The results of quality control (QC) tests on 70 mammography units in Southern California from 1986 to 1990 are reported. Thirteen facilities, selected because they housed all of the mammography units in three communities involved in a National Institutes of Health research project, had their units tested twice at an interval of 1 year. Fourteen self-selected units were also tested twice at intervals ranging from 1-3 years. Forty-three self-selected units in 31 additional facilities had testing only once. All 70 units underwent measurement of focal spot size or resolution, tube output, half-value layer (HVL), automatic exposure control (AEC) accuracy, relative kVp accuracy, mean glandular dose, and imaging of several test objects. The test results for the units tested once showed no significant differences compared to those tested twice. For the latter, once the units were tested and determined to be acceptable, retesting showed differences only in overall film optical density, dose, and AEC performance.

Establishing minimum performance standards, calibration intervals, and optimal exposure values for a whole breast digital mammography unit.

Methods are developed to establish minimum performance standards, calibration intervals, and criteria for exposure control for a whole breast digital mammography system. A prototype phantom was designed, and an automatic method programmed, to analyze CNR, resolution, and dynamic range between CCD components in the image receptor and over time. The phantom was imaged over a 5 month period and the results are analyzed to predict future performance. White field recalibration was analyzed by subtracting white fields obtained at different intervals. Exposure effects were compared by imaging the prototype phantom at different kVp, filtration (Mo vs Rh) and mAs. Calcification detection tests showed that phantom images, obtained at 28 kVp with a Mo/Mo anode/filter and low mAs technique, often could not depict Al2O3 specks 0.24 mm in diameter, while a 28 kVp Mo/Rh, higher mAs technique usually could. Stability of the system tested suggests that monthly phantom imaging may suffice. Differences in CCD performance are greater (12%) than differences in a single CCD over time (6%). White field recalibration is needed weekly because of pixel variations in sensitivity which occur if longer intervals between recalibration occur. When mean glandular dose is matched, Rh filtration gives better phantom performance at 28 kVp than Mo filtration at 26 kVp and is recommended for clinical exposures. An aluminum step wedge shows markedly increased dynamic range when exit exposure is increased by using a higher energy spectrum beam. Phantoms for digital mammography units should cover the entire image receptor, should test intersections between components of the receptor, and should be automatically analyzed.

Recommendations on performance characteristics of diagnostic exposure meters: report of AAPM Diagnostic X-Ray Imaging Task Group No. 6.

Task Group 6 of the Diagnostic X-Ray Imaging Committee of the American Association of Physicists in Medicine (AAPM) was appointed to develop performance standards for diagnostic x-ray exposure meters. The recommendations as approved by the Diagnostic X-Ray Imaging Committee and the Science Council of the AAPM are delineated in this report and provide specifications on meter precision, calibration accuracy, calibration reference points, linearity, energy dependence, exposure rate dependence, leakage, amplification gain settings, directional dependence, the stem effect, constancy checks, and calibration intervals. The report summarizes recommendations for meters used in mammography, general purpose radiography including special procedures, computed tomography, and radiation safety surveys for x-ray radiography.

New and future developments in screen-film mammography equipment and techniques.

Two improvements in mammography equipment during the last 4 years will greatly affect the practice of mammography during the next decade: dose reduction and improved testing equipment for monitoring the quality of mammography. The increased use of digital radiography has stimulated studies comparing digital mammography with screen-film mammography. Some digital algorithms for detecting clusters of calcifications may have application in screening mammography within the next decade.

The relative effects of system parameters on texture in gray-scale ultrasonograms.

Texture in B-mode ultrasonography of the abdomen can, in many cases, serve as an important diagnostic tool. Unfortunately, because texture as it is now presented depends on many parameters which are independent of pathology, it is not always a reliable sign. Parameters such as transducer size, center frequency, and bandwidth; B-mode display and storage methods; near and far field wave effects; and the type of intervening tissue all influence the display of texture. We have investigated these parameters of texture through the analysis of computer simulated B-mode ultrasonograms based on in vitro A-line data collected under a variety of circumstances and conditions. A freshly excised bovine pancreas served as the target tissue. Computer generated textured B-mode images, produced with different gray-scale and scan converter storage algorithms, were evaluated by a panel of six radiologists who individually divided sixty images into groups of images with similar texture. These evaluations were mapped into a six feature hyperspace with each axis representing a particular parameter. By applying a variation of the principal components transformation, each parameter could be ranked according to its effect on texture. We found that the type of gray-level preprocessing map had the most effect on texture; intervening tissue and transducer variations had intermediate effects. So long as the target distance did not exceed the focal length of the transducer, the distance between the transducer and the target tissues had the least effect on texture.

Ultrasound breast examination of asymptomatic patients with normal but radiodense mammograms.

Of 1757 consecutive women patients examined with ultrasound (US) between 1983-1986, 796 patients were asymptomatic, had no masses on their mammograms, and were scanned with whole breast US because their breasts were radiodense on mammography. Combinations of automated and real-time hand-held equipment, ranging in frequency from 3.9 MHz to 10 MHz. were used. Of the 796 examinations, 71.5%(569) were normal, 23.5%(187) disclosed cysts, and 4.3%(40) revealed solid masses. Four masses were biopsied, yielding one malignancy, two fibroadenomas and one normal breast tissue. All but 4 of the remaining 36 asymptomatic women with solid masses had a follow-up period of at least 1 year. In the authors' opinion, this low yield of one breast cancer in 796 asymptomatic women implies that whole breast US is of little value (in terms of breast cancer detection) in breasts that are radiodense but otherwise negative on mammography.