Digital mammography: observer performance study of the effects of pixel size on the characterization of malignant and benign microcalcifications.

RATIONALE AND OBJECTIVES: The authors performed this study to evaluate the effects of pixel size on the characterization of mammographic microcalcifications by radiologists. MATERIALS AND METHODS: Two-view mammograms of 112 microcalcification clusters were digitized with a laser scanner at a pixel size of 35 microm. Images with pixel sizes of 70, 105, and 140 microm were derived from the 35-microm-pixel size images by averaging neighboring pixels. The malignancy or benignity of the microcalcifications had been determined with findings at biopsy or 2-year follow-up. Region-of-interest images containing the microcalcifications were printed with a laser imager. Seven radiologists participated in a receiver operating characteristic (ROC) study to estimate the likelihood of malignancy. The classification accuracy was quantified with the area under the ROC curve (Az). The statistical significance of the differences in the Az values for different pixel sizes was estimated with the Dorfman-Berbaum-Metz method and the Student paired t test. The variance components were analyzed with a bootstrap method. RESULTS: The higher-resolution images did not result in better classification; the average Az with a pixel size of 35 microm was lower than that with pixel sizes of 70 and 105 microm. The differences in Az between different pixel sizes did not achieve statistical significance. CONCLUSION: Pixel sizes in the range studied do not have a strong effect on radiologists' accuracy in the characterization of microcalcifications. The low specificity of the image features of microcalcifications and the large interobserver and intraobserver variabilities may have prevented small advantages in image resolution from being observed.

Discrimination of sonographically detected breast masses using frequency shift color Doppler imaging in combination with age and gray scale criteria.

Frequency shift color Doppler imaging was assessed in conjunction with patient age and gray scale (GS) features for discriminating benign from malignant breast masses. Thirty-eight women with sonographically detected masses scheduled for biopsy were evaluated using a 6- to 13-MHz scan head, and the masses were delineated in ultrasonographic image volumes. Vascularity in and around each mass was quantified using speed-weighted pixel density (SWD). Gray scale features were ranked visually on a linear scale. Combinations of indices were compared with histologic findings (18 benign and 20 malignant). Receiver operating characteristic analysis ranked performance in decreasing order from the SWD-Age-GS index, to SWD-GS, SWD-Age, Age-GS, GS criteria, SWD, and Age. At 100% sensitivity, SWD-Age-GS, SWD-GS, and SWD-Age discriminated benign from malignant masses with specificities of 94%, 89%, and 72%, respectively. These results indicate significant improvement in ultrasonographic discrimination of sonographically detected breast masses by combining the vascularity measure SWD with age and GS criteria.

Adverse effects of increased body weight on quantitative measures of mammographic image quality.

OBJECTIVE: The purpose of our study was to show that compressed breast thickness on mammograms in overweight and obese women exceeds the thickness in normal-weight women and that increased thickness results in image degradation. SUBJECTS AND METHODS: Three hundred consecutive routine mammograms were reviewed. Patients were categorized according to body mass index. Compression thickness, compressive force, kilovoltage, and milliampere-seconds were recorded. Geometric unsharpness and contrast degradation were calculated for each body mass index category. RESULTS: Body mass index categories were lean (3%), normal (36%), overweight (36%), and obese (25%). Body mass index was directly correlated with compressed thickness. In the mediolateral oblique view, the mean thickness of the obese category exceeded normal thickness by 18 mm (p < 0.01), corresponding to a 32% increase in geometric unsharpness. Mean obese thickness exceeded lean thickness by 33 mm (p < 0.01), corresponding to a 79% increase in unsharpness. Similar trends were observed for the craniocaudal view. In the mediolateral oblique projection, there was an increase of 1.0 kVp (p < 0.01) for obese compared with normal and 1.7 kVp (p < 0.01) between lean and obese, corresponding, respectively, to a 16% and a 25% decrease in image contrast because of scatter and kilovoltage changes. Milliampere-seconds increased by 47% on the mediolateral oblique images in the obese category compared with normal body mass index. CONCLUSION: An increased body mass index was associated with greater compressed breast thickness, resulting in increased geometric unsharpness, decreased image contrast, and greater potential for motion unsharpness.

Improvement of radiologists' characterization of mammographic masses by using computer-aided diagnosis: an ROC study.

PURPOSE: To evaluate the effects of computer-aided diagnosis (CAD) on radiologists' classification of malignant and benign masses seen on mammograms. MATERIALS AND METHODS: The authors previously developed an automated computer program for estimation of the relative malignancy rating of masses. In the present study, the authors conducted observer performance experiments with receiver operating characteristic (ROC) methodology to evaluate the effects of computer estimates on radiologists' confidence ratings. Six radiologists assessed biopsy-proved masses with and without CAD. Two experiments, one with a single view and the other with two views, were conducted. The classification accuracy was quantified by using the area under the ROC curve, Az. RESULTS: For the reading of 238 images, the Az value for the computer classifier was 0.92. The radiologists' Az values ranged from 0.79 to 0.92 without CAD and improved to 0.87-0.96 with CAD. For the reading of a subset of 76 paired views, the radiologists' Az values ranged from 0.88 to 0.95 without CAD and improved to 0.93-0.97 with CAD. Improvements in the reading of the two sets of images were statistically significant (P = .022 and .007, respectively). An improved positive predictive value as a function of the false-negative fraction was predicted from the improved ROC curves. CONCLUSION: CAD may be useful for assisting radiologists in classification of masses and thereby potentially help reduce unnecessary biopsies.

Imaging of nontraumatic hemorrhage of the adrenal gland.

Nontraumatic hemorrhage of the adrenal gland is uncommon. The causes of such hemorrhage can be classified into five categories: (a) stress, (b) hemorrhagic diathesis or coagulopathy, (c) neonatal stress, (d) underlying adrenal tumors, and (e) idiopathic disease. Computed tomography (CT), ultrasonography (US), and magnetic resonance (MR) imaging play an important role in diagnosis and management. CT is the modality of choice for evaluation of adrenal hemorrhage in a patient with a history of stress or a hemorrhagic diathesis or coagulopathy (anticoagulant therapy). CT may yield the first clue to the diagnosis of adrenal insufficiency secondary to bilateral massive adrenal hemorrhage; such insufficiency is rare but life threatening. US is the modality of choice for evaluation of neonatal hematoma, and MR imaging is helpful for further characterization. MR imaging is also useful in the diagnosis of coexistent renal vein thrombosis. When an adrenal abscess is suspected, percutaneous aspiration and drainage under imaging guidance should be performed. Hemorrhage into an adrenal cyst or tumor can cause acute onset of symptoms and signs in a patient without discernible risk factors for adrenal hemorrhage. A hemorrhagic adrenal tumor should be suspected when CT or MR imaging reveals a hemorrhagic adrenal mass of heterogeneous attenuation or signal intensity that demonstrates enhancement.

Semiautomatic registration of volumetric ultrasound scans.

We demonstrate the ability to register easily and accurately volumetric ultrasound scans without significant data preprocessing or user intervention. Two volumetric ultrasound breast scan data sets were acquired from two different patients with breast cancer. Volumetric scan data were acquired by manually sweeping a linear array transducer mounted on a linear slider with a position encoder. The volumetric data set pairs consisted of color flow and/or power mode Doppler data sets acquired serially on the same patients. A previously described semiautomatic registration method based on maximizing mutual information was used to determine the transform between data sets. The results suggest that, even for the deformable breast, three-dimensional full affine transforms can be sufficient to obtain clinically useful registrations; warping may be necessary for increased registration accuracy. In conclusion, mutual information-based automatic registration as implemented on modern workstations is capable of yielding clinically useful registrations in times <35 min.

Relationship between mammographic screening intervals and size and histology of ductal carcinoma in situ.

OBJECTIVE: The objective of this study was to determine how the length of time between mammographic screenings is related to the size, grade, and histology of mammographically detected ductal carcinoma in situ (DCIS). MATERIALS AND METHODS: We retrospectively reviewed 166 consecutive mammograms of women evaluated for DCIS with (n = 24) and without (n = 142) microinvasion. The size of the DCIS was determined by the maximum diameter as measured on the mammogram. After pathologic analysis, DCIS was classified by histologic architecture, nuclear grade, presence of microinvasion, and presence of multifocality. Four screening intervals were defined: annual (6-17 months), biennial (18-29 months), triennial (> or = 30 months), and first time. Patients were grouped according to screening intervals. The average age of all groups was 55 years. RESULTS: The annual group (mean size of DCIS, 1.69 cm) had significantly smaller DCIS than did the biennial (mean size, 2.27 cm), triennial (mean size, 3.49 cm), or first time groups (mean size, 3.29 cm) (p = .003). Comedo histology was more frequently observed in patients screened biennially (73.7%) than in those screened annually (46.8%) (p = .05). High-grade nuclear histology was more commonly seen in the biennial (76.3%) than in the annual (48.1%) screening group (p = .008). We found no significant correlation between screening interval and the incidence of microinvasion and multifocality. CONCLUSION: Small, low-grade noncomedo DCIS was more common in the annual mammographic screening group than in the biennial screening group. A direct relationship was found between DCIS size and length of screening interval: DCIS detected at annual screening was smaller than that found at biennial screening, which in turn was smaller than DCIS revealed at triennial screening. This study provides inferential support for annual screening mammography for DCIS detection and management.

Limitations of student evaluations of curriculum.

RATIONALE AND OBJECTIVES: Medical student surveys are used extensively in the development and modification of curriculum. The purpose of this study was to look at medical student surveys of a radiology lecture series, evaluating the accuracy of student perceptions of learning and factors affecting them. MATERIALS AND METHODS: After a "Case of the Week" lecture series, 156 3rd-year medical students returned a survey evaluating the experience with 10 questions on a four-point scale (1 = disagree, 4 = agree very much) and took a clinical competency assessment (CCA) examination with a radiology substation. Survey responses were compared with actual examination performance, analyzed for how overall learning was characterized in specific educational objectives, and evaluated for factors affecting perceived learning. RESULTS: The mean response for perceived CCA examination preparedness was 1.83. The mean radiology station test score was 90.43%. Correlations between student perception of learning and the scoring of focused learning objectives ranged from 0.33 to 0.48 (P < .01). Students responding 1 to items assessing perceived lecture organization, stimulation to read, and interest in the field of radiology had mean scores for perception of overall learning of 2.09-2.44 and mean scores for recommendation of course continuation of 1.68-2.46. Students responding 4 had means of 3.25-3.81 and 3.06-4.0, respectively. CONCLUSION: Student perceptions of the value of curriculum were inaccurate compared with external measures of performance, and students poorly related their general impressions to specific learning objectives. Perceived lecture organization, stimulation to read, and interest in radiology as a specialty affected perceived overall learning and perceived value of the lecture series.

Mammographic appearance of recurrent breast carcinoma in six patients with TRAM flap breast reconstructions.

PURPOSE: To determine the mammographic appearance of locally recurrent cancer in patients with breast reconstructions with transverse rectus abdominis musculocutaneous (TRAM) flaps after mastectomy for primary breast cancer. MATERIALS AND METHODS: The mammograms and records of women treated for breast cancer with mastectomy and TRAM flap reconstruction who developed local recurrences from 1995 to 1997 were reviewed retrospectively. Eight cancers were identified in six women. Five women had palpable abnormalities, and the sixth had cancer detected at screening mammography at another institution. Mean age at recurrence was 48 years. RESULTS: All eight cancers were visible on mammograms: four masses, two pleomorphic microcalcifications, and two masses with calcifications. Four of the eight were in the upper central portion of the breast. Before reconstruction, the original histologic diagnosis for all cases had been multifocal ductal carcinoma in situ. All recurrences were invasive cancer. Median time from the original diagnosis of breast cancer to diagnosis of recurrence was 42 months. Two of four patients who subsequently underwent axillary node dissection had metastatic disease in the lymph nodes. The single patient who underwent mammographic screening (elsewhere) had negative axillary lymph nodes. CONCLUSION: The mammographic appearance of recurrent carcinoma in TRAM flap reconstructions is similar to that of primary breast cancer.

Patient satisfaction with screening mammography: online vs off-line interpretation.

RATIONALE AND OBJECTIVES: The purpose of this study was to determine patient satisfaction with online and off-line methods of interpreting screening mammograms and to determine those characteristics that most influence patient preferences. MATERIALS AND METHODS: A survey consisting of 17 questions assessing satisfaction and demographics was distributed to women whose screening mammograms were read in one of two ways at the same facility: (a) the "online" method, in which all additional requested studies are completed and results are immediately communicated, and (b) the "off-line" method, in which mammograms are read in batches the next day, results are communicated at a later time, and patients return for diagnostic studies. To ensure statistical power, at least 174 surveys were collected from each group. RESULTS: Of 192 women whose mammograms were interpreted online, 188 (97%) responded, and of 247 women whose mammograms were interpreted off-line, 181 (73%) responded. Overall mean satisfaction with the mammographic experience and with the time it took to receive results differed significantly between the two groups (P < .002). Analysis of the subgroup of very satisfied respondents disclosed that 92% of patients in the online group and 59% of patients in the off-line group were very satisfied with their mammography experience. If allowed to choose the method of interpretation, 97% of patients in the online group and 91% of patients in the off-line group would choose online interpretation. CONCLUSION: Patients preferred online mammographic interpretation of screening studies.

3-D color Doppler image quantification of breast masses.

In this article, new measures obtained from color Doppler images are introduced and a pilot study is described, in which these and previously published indices are evaluated for use in future work. Twenty women with breast masses observed on mammography and going to surgical biopsy were studied. Of the masses, 11 proved to be benign and 9 were malignant. Both 3-D mean frequency shift (f-CDI) and power mode Doppler (p-CDI) imaging were performed. To identify the mass and other regions of interest, vessels were displayed as rotatable 3-D color volumes, superimposed on selectable grey-scale/color flow slices. Doppler signals were recorded in each of 6 ellipsoidal regions of interest in and around the mass and 2 in normal tissues. Seven measures were computed in each region, three from power mode, two from mean frequency and two from combinations of both. Radiologists rated the grey-scale appearances of the masses on a scale of 1 to 5 (5=most suspicious) for each of 6 conventional grey-scale criteria. Of the individual vascularity measures in individual ROIs, the log speed-weighted pixel density and log power-weighted pixel density in the lesion internal periphery showed the greatest discrimination of malignancy, although neither was statistically significant nor as good as the peak variables described below. The mean visual grey-scale rating was the best discriminator overall, but two peak vascularity measures each made promising scatterplots in conjunction with the average visual grey-scale rating. These two vascularity measures were the log peak normalized power-weighted pixel density (peak NPD) and log of peak mean Doppler frequency times the peak NPD (vM x NPD(M)). Each of these two values was the maximum in any one of the five chosen ROIs closely associated with the mass. A possible rationale for the relative success of these peak values is the blood signal's normalization and the inhomogeneity of most breast cancers and the expectation that the highest velocities (shunting) and largest collections of blood are not necessarily in the same region in and around the tumor. Peak NPD of cancers varied with age, decreasing by a factor of 45 from 33 to 77 y.

Breast cancer in women who undergo screening mammography: relationship of hormone replacement therapy to stage and detection method.

PURPOSE: To compare the breast cancer stages and detection methods in screened women who receive hormone replacement therapy (HRT) with those in screened women who do not receive HRT to determine whether HRT affects the stage or mammographic detection of malignancy. MATERIALS AND METHODS: One hundred fifteen cases of breast cancer in women (age range, 55-65 years) in whom at least one screening mammogram had been obtained at least 24 months before diagnosis and in whom the history regarding HRT could be determined were reviewed retrospectively. Statistical analysis was performed with CHI-2 analysis and the Fischer exact test. RESULTS: The cancer stages in the 58 women who received HRT were stage 0 in 15 (26%), stage I in 28 (48%), stage II in 13 (22%), and stage III in two (3%) women. The stages in the 57 women who did not receive HRT were stage 0 in 19 (33%), stage I in 24 (42%), stage II in 11 (19%), stage III in two (4%), and stage IV in one (2%) woman. Cancers in 38 (67%) of the women who did not receive HRT and in 38 (66%) of those who did receive HRT were detected with mammography alone; false-negative mammograms were obtained in five (9%) women in the non-HRT group and in four (7%) women in the HRT group (P = .89). CONCLUSION: Among screened women who developed breast cancer, there were no significant differences in cancer stages or in the number of mammographically detected cancers or false-negative mammograms between the HRT group and the non-HRT group.

Value of case-of-the-week presentations on the radiology component of the Comprehensive Clinical Assessment examination.

RATIONALE AND OBJECTIVES: The Comprehensive Clinical Assessment (CCA) examination at the University of Michigan Medical School is a series of test stations through which the mastery of clinical skills is evaluated. The purpose of this study was to determine whether student performance on the radiology station improved in years when radiology faculty presented case-of-the-week unknowns to the 3rd-year students. MATERIALS AND METHODS: The authors compared four separate classes of medical students in examination years 1993, 1994, 1995, and 1996 by using the total CCA examination score, the radiology station score, and radiology station pass/fail rates. Radiology case-of-the-week presentations were given by the radiology faculty only in academic years 1993-1994 and 1994-1995 (examination years 1994 and 1995). RESULTS: The means and standard deviations of the radiology station scores for the examination years when case-of-the-week presentations were not given, 1993 and 1996, were 78.92 +/- 13.62 and 79.76 +/- 13.62, respectively. In the years case-of-the-week presentations were given, 1994 and 1995, the radiology station scores averaged 90.83 +/- 8.58 and 89.97 +/- 9.66, respectively (P < .001, global alpha = .05). Total CCA percentage correct scores were similar for all years studied. In 1993 and 1996, 7.6% and 5.3% of students, respectively, failed the radiology station. In 1994 and 1995, 0.4% and 0% of students, respectively, failed (P < .0001). CONCLUSION: Case-of-the-week presentations by radiology faculty increased 3rd-year students' basic radiologic knowledge as evidenced by increased scores on the radiology station of the CCA examination.

Mammographic appearance of breast cancer in African-American women: report of 100 consecutive cases.

RATIONALE AND OBJECTIVES: The authors determined the mammographic appearance of breast carcinoma in African-American women and compared it with that in a white cohort. MATERIALS AND METHODS: The authors reviewed the mammograms, clinical records, and pathology records of 97 consecutive African-American women with 100 confirmed breast cancers and 110 white women with 111 confirmed breast cancers. RESULTS: The mammograms obtained in African-American women were positive in 94 cases (94%), and those obtained in white women were positive in 99 cases (89%). Forty-seven percent of malignancies in African-American women appeared as calcifications, alone or with a mass, and 41% appeared as a mass only. There was no statistically significant difference in the frequency of these two findings between the African-American and the white populations. There was no statistically significant difference in the breast parenchymal pattern between the two groups. The most common tumor location in both races was the upper outer quadrant. CONCLUSION: Breast carcinoma in African-American women is similar to that in white women in terms of mammographic appearance, location, and breast density. The mammographic appearance should not be an impediment to the detection of breast cancer in African-American women.

Metaplastic carcinoma of the breast: mammographic appearance with pathologic correlation.

OBJECTIVE: The objective of this study was to describe the mammographic appearance with pathologic correlation of metaplastic carcinoma of the breast. CONCLUSION: Metaplastic carcinomas of the breast are masses with mammographic characteristics of low suspicion because of their histologic appearance. Metaplastic carcinoma should be included in the differential assessment of predominately circumscribed, noncalcified masses revealed on mammography. One salient feature that may distinguish metaplastic carcinomas is the occurrence of a circumscribed portion with a spiculated portion, which is seen in carcinomas that have a significant mixture of metaplastic and invasive carcinoma growth patterns.

Women with breast cancer: histologic findings in the contralateral breast.

PURPOSE: To investigate contralateral breast biopsy histologic findings in women with breast cancer. MATERIALS AND METHODS: Histologic findings in 237 patients with breast cancer who underwent contralateral breast biopsy for clinically or mammographically detected abnormalities were retrospectively reviewed. Malignant findings were categorized by histologic type. Benign findings were categorized by risk of breast cancer. Comparison was made with mammographically guided breast biopsy results in 1,294 patients without breast cancer. RESULTS: Of the 237 patients, 168 (70.9%) had either malignancy or high-risk histologic findings. One hundred thirty-nine patients (58.6%) had malignant findings; 98 (41.4%) had benign findings. Of the 98 with benign findings, 29 (30%) had high-risk histologic findings. Thirty (33%) of the 91 patients with invasive cancer had invasive lobular carcinoma. Forty-seven (45.6%) of the 103 patients with malignant lesions at mammographically guided biopsies had ductal carcinoma in situ alone. CONCLUSION: Compared with biopsy in women without breast cancer, contralateral biopsy in women with breast cancer was more likely to show malignancy, invasive lobular carcinoma, or ductal carcinoma in situ alone (P < .001) or to show high-risk histologic benign findings (P < .001). Mammographic and clinical findings in the contralateral breast should be regarded as more suspicious than those in patients without known breast cancer.

Mammography of breast carcinoma in women who have mutations of the breast cancer gene BRCA1: initial experience.

OBJECTIVE: We describe for the first time (to our knowledge) the mammographic appearance of breast cancer and breast density in carriers of the breast cancer gene BRCA1. CONCLUSION: The mammographic appearance of breast cancer in BRCA1 carriers was similar to that in the general population. All mammographic densities were observed. Appropriate mammographic management of these extremely high-risk patients is yet to be determined.

The 3D and 2D color flow display of breast masses.

A prospective study was performed in 24 women with breast masses on mammography going on to surgical biopsy. 2D and 3D power mode and frequency shift color flow Doppler scanning and display were compared. Vessels were displayed as rotatable color volumes in 3D, superimposed on gray-scale slices. The latter were stepped sequentially through the imaged volume. Radiologists rated the masses in each display (3D, 2D and videotapes) on a scale of 1 to 5 (5 = most suspicious) for each of six conventional gray-scale and six new vascular criteria. Thirteen masses proved to be benign and 11 were malignant. 3D provided a stronger subjective appreciation of vascular morphology and allowed somewhat better ultrasound discrimination of malignant masses than did the 2D images or videotapes (specificities of 85%, 79% and 71%, respectively, at a sensitivity of 90%). Only in 3D did the vascularity measures display a trend towards significance in this small study.