Figure 1. Film gradient.Äëÿ ó÷àñòíèêîâ
áåñåäû î ñêðèíèíãå ðàêà ìîëî÷íîé æåëåçû.
Many exciting developments are occurring in breast imaging. Some, such as digital mammography, have the potential to change the field radically, while others will have effects that are more subtle but are equally important. Digital mammography holds the promise of telemammography and computer-aided diagnosis. Mammoscintigraphy may be helpful in identifying drug-resistant tumors before therapy. There is renewed interest in evaluating ultrasound as a potential adjunctive screening tool in women with radiographically dense breasts. Finally, contrast-enhanced magnetic resonance (MR) imaging may be used more extensively in monitoring tumor response to primary chemotherapy and in the preoperative assessment of patients being considered for breast conservation therapy.
Mammography remains one of the few areas of medical imaging yet to make the transition to a digital format. However, several manufacturers of mammography equipment have digital systems that will soon become commercially available. [11] Digital mammography offers many potential advantages over current film-screen techniques, but there are challenges associated with this new technology that must be overcome before all its potential benefits are realized.
In film-screen mammography, x-ray photons that pass through the breast interact with a phosphor screen causing it to emit light photons that expose an adjacent sheet of film and produce a latent image. This latent image is made visible by film processing and is viewed on a high luminance viewbox. Once the image is produced, it cannot be manipulated in any way. In digital mammography, the x-ray photons interact with a detector that emits digital signals that are electronically processed into an image. This image can then be viewed on a high resolution, high luminance monitor at a computer workstation or printed onto film. An important difference between film-screen and digital systems is that the digital image can be manipulated after it has been acquired (post-processing). Adjustments in the brightness and contrast levels are easily accomplished, as is electronic magnification of selected regions. This postprocessing manipulation may allow the detection of subtle abnormalities within dense tissue. Furthermore, the ability to perform electronic magnification may reduce the need to perform additional magnification views, each of which exposes the patient to additional radiation. [18]
Wide Exposure LatitudeThe ability to discern discrete structures (either normal or abnormal) within the breast results from the differential x-ray-attenuating characteristics of these structures, that is, the subject contrast. For example, fat attenuates to a much lesser degree than does fibroglandular tissue, and these two substances appear very different on the x-ray image. However, the final displayed contrast of the film image also depends on the imaging characteristics of the film, the film gradient. The response of the film to light exposure from the phosphor screen is described by a sigmoidal curve (Fig. 1) . Maximum contrast occurs in the central, linear portion of the curve where the change in film optical density for a given change in exposure is greatest. Thus, this is the range in which structures that differ very little in attenuating characteristics (e.g., fibroglandular tissue and masses) would most likely be displayed as different densities on film. The range of exposures that fall within this optimum zone (exposure latitude) is limited, no more than 40:1 for most film/screen combinations. [57] The range of x-ray photons exiting the breast (that are converted to light photons in the phosphor screen), from the very dense fibroglandular to the very fatty portions, however, can be more than 100:1. [57] This results in imaging in the toe or shoulder portions of the
In contrast, digital detectors exhibit a linear, rather than a sigmoidal, response over exposure ranges of over 1000:1 and thus have much wider exposure latitudes. [57] With these systems, optimum display contrast is achieved in both the highly attenuating (dense fibroglandular) and less attenuating (fatty) portions of the breast. Postprocessing allows the image to be further optimized for evaluating specific areas.
Computer-Aided DiagnosisComputer programs to aid in the detection and characterization of breast masses and calcifications have been developed and tested. These programs are designed not only to identify and flag areas of potential
Studies to date have primarily utilized film mammograms that were digitized by means of a film digitizer; however, in the era of direct digital mammography, these systems could be incorporated into mammography machines so that the computer could directly analyze the digital image as it is obtained. The radiologist at the workstation could then refer to the results of this computer analysis following his or her own review of the mammogram and render a final opinion.
Image Storage and TransmissionIn film-screen mammography, the x-ray film not only functions as an image acquisition and display device but as a storage device as well. If it is ever lost or damaged, the information it holds may be lost forever. In digital mammography these three functions are separated. As mentioned above, the image is acquired via a digital detector and displayed by a computer monitor. Efficient long-term storage can be performed by optical disks or other electronic storage devices.
The digital format lends itself well to telemammography which allows the efficient transfer of images between institutions to facilitate patient care and eliminates the problems of poor quality copy films or lost original films. Telemammography may also bring the expertise of leading mammographers to even the most remote locations for on-line consultation. [18]
Low Image NoiseImage noise resulting from film granularity is a cause of degradation of the film image. This problem is more pronounced in faster film-screen systems but exists to a certain degree in all systems. It is eliminated in digital mammography read on a computer monitor (soft copy).
Digital mammography may be quicker than film-screen mammography. Processing x-ray films following the mammographic procedure may add more than 5 minutes to the total examination time. Digital images are available almost instantaneously following exposure, so screening more patients/day may be possible. [18]
One of the main limitations of digital mammography is the reduced spatial resolution of digital detectors compared with film-screen mammography. Spatial resolution refers to the ability of an imaging system to display, separately, two small objects placed close together. One common method of measuring spatial resolution is the use of a test object consisting of thin strips of metal of equal thickness separated by a distance equal to the thickness of strips. Each strip and its adjoining space is called a line pair (lp). Screen-film mammographic systems have a theoretical spatial resolution of up to 20 lp/mm when measured with high-contrast test objects. This resolution could conceivably allow the display of objects as small as 25 mum (1 mm/40). However, because of the inherent noise and limited exposure latitude of film mammography, this resolution is not typically achieved in clinical imaging. It has been shown that calcifications less than 130 mum in size are rarely detected on film-screen mammography. [23] The actual minimum spatial resolution required for high quality mammography is unknown.
A digital image is composed of an array of discrete picture elements, or pixels. The smaller these pixels are, the greater the resolving power of the system and the sharper the image. Digital detectors that can image at a 50 mum pixel size currently exist, [11] equivalent to a resolution of 10 lp/mm. Although this is less than the resolution of film-screen mammography, the lower intrinsic noise and better contrast characteristics of digital detectors may result in equivalent clinical performance. [57]
DisplayThe highest resolution computer monitors that are now commercially available have approximately 2000 by 2000 pixels. This size is insufficient to display an image of the entire breast at a 50 mum pixel size (10 lp/mm). Such a task would require a monitor with 4000 by 5000 pixels. These monitors are extremely expensive and are not commercially available. Soft-copy interpretation of mammographic images would
The three leading manufacturers involved in digital mammography technology have each chosen a slightly different approach to the design of this equipment. TREX Medical (Danbury, CT) uses a cesium iodide detector that is connected to a mosaic of charged-coupled devices (CCDs) via fiberoptic tapers. [11] [58] When exposed to the transmitted radiation exiting the breast, the cesium iodide detector scintillates. The resulting light photons are transferred to the CCDs via the fiberoptic tapers where they are converted to the digital signals that make up the image. The device made by GE Medical Systems (Schenectady, NY) uses a flat panel of amorphous silicon within which is an array of photodiodes. [11] [58] The scintillator is a coating of cesium iodide on the surface of the photodiodes. Here the photodiodes covert the light photons, produced by the cesium iodide, into the digital image. Fischer Imaging (Boulder, CO) uses a slot-scanning system in which the x-ray beam, confined to an area 24 cm by 1.4 cm, scans the breast from side to side. [11] This narrow beam is focused on, and moves in concert with, a similar sized detector on the x-ray exit surface of the breast. This detector is cesium iodide that is connected to CCDs via fiberoptic tapers.
Radionuclide imaging techniques that have been applied to the breast include fluorine-18-fluorodeoxyglucose positron emission tomography (FDG PET), and scintigraphy with thallium Tl 201 or technetium Tc 99m sestamibi. The high cost and limited availability of FDG PET limit its usefulness today. A high sensitivity in the detection of breast carcinoma has been reported with thallium Tl 201 [56] ; however, thallium Tl 201 is not an ideal imaging agent because of its gamma emission characteristics and the absorbed dose. The imaging agent that has received the most interest recently is technetium Tc 99m sestamibi. This was approved by the Food and Drug Administration in 1991 as a myocardial imaging agent (Cardiolite, Dupont Merck Pharmaceutical,
The sensitivity of technetium Tc 99m sestamibi in the detection of breast cancer is highly dependent on tumor size. Sensitivity rates of 0.92 to 1.0 have been reported for palpable lesions (Table 1) ; for nonpalpable lesions, rates as low as 0.25 have been observed (Table 2) . Scopinaro et al recently reported sensitivities of 0.26, 0.56, 0.95 and 0.97 for T1a, T1b, T1c and T2 breast cancers respectively. [48] This poor performance of mammoscintigraphy with small tumors results, in large part, from the design of available gamma cameras. Development of dedicated, small field-of-view instruments is currently underway. These are likely to improve significantly the detection rate of small tumors.
The low sensitivity of mammoscintigraphy in the detection of nonpalpable
breast cancer makes it unsuitable as a screening modality.
Includes 40 nonpalpable
lesions detected on mammography. Results for this
subset not provided by authors.It has been suggested that technetium Tc 99m mammoscintigraphy can reduce the number of unnecessary breast biopsies by further characterizing mammographic or palpable abnormalities. For mammographic (nonpalpable) abnormalities, this is unfeasible. In fact, Tolmos et al, reflecting on the low sensitivity of the technique for nonpalpable breast cancer, have stated that "Scintimammography, at its current stage of development, is not an appropriate second screening test to determine whether or not a patient with an abnormal mammogram ... should have a breast biopsy." [52] It is also unlikely that a negative scan will deter biopsy of clinically suspicious palpable masses, given medicolegal considerations.
Technetium Tc 99m sestamibi mammoscintigraphy has also shown some efficacy in detecting metastatic involvement of axillary lymph nodes. [51] [53] This use may become less relevant as the technique of sentinel lymphadenectomy becomes more popular, as sentinel lymphadenectomy provides a highly reliable means of assessing the status of the axilla that is less invasive than a complete axillary dissection.
Ultrasound is an important tool in the practice of modern breast imaging. Well established indications for breast ultrasound include:
There has been renewed interest recently in whether ultrasound can have a more substantial effect on breast imaging, and specifically, whether ultrasound can mitigate some of the weaknesses of mammography, namely, its reduced sensitivity in radiographically dense breasts (leading to missed cancers) and low positive predictive value (resulting in unnecessary biopsies).
Early clinical trials of breast ultrasound consistently showed that it was unsuitable for screening because of the low rate of detection of unsuspected malignancies in patients with normal mammograms and clinical breast examinations. [2] [10] [9] [28] These trials suffered from the limitations of the technology available in the early to mid-1980s and from the absence, in most of them, of direct physician scanning.
There have been tremendous improvements in ultrasound technology since those early studies were performed, including the availability of high frequency, hand-held probes that deliver enhanced spatial resolution and the ability to image in real time. Some recent studies lend support to the belief that modern ultrasound may be capable of better performance than earlier equipment.
Gordon and Goldenberg reported retrospectively on the results of 12,706 ultrasound examinations performed over a 5-year period to evaluate a palpable or mammographic abnormality. [15] In every case, after targeted ultrasound of the area of concern was performed, a survey of the remainder of the ipsilateral breast was carried out. ultrasound scanning was performed by a radiologist. Survey ultrasound noted 1575 incidental masses. Of these, 1296 were judged to be of low suspicion, and the patients were advised to undergo follow-up ultrasound at 6, 12, 24 and 36 months. The remaining 279 masses underwent ultrasound-guided fine-needle aspiration cytology and, among these, 44 malignancies in 30 patients were eventually confirmed surgically. Of these, 16 masses in 15 patients represented additional foci of malignancy in patients
Kolb and co-workers recently reported the results of a prospective trial of physician-performed screening ultrasound. [27] Of 11,220 consecutive patients, 3626 women had radiographically dense breasts and normal mammograms and clinical breast examinations. All these women underwent screening ultrasound. In 215 patients, 279 solid masses were identified. There was a statistically significant linear increase in the likelihood of finding an incidental mass with increasing breast density. Of these 215 patients, 111 underwent fine-needle aspiration cytology, 92 underwent follow-up ultrasound (86 stable after at least one follow-up at 6 months), and 12 underwent surgical biopsy based on the ultrasound findings. In the group receiving fine-needle aspiration, 11 cancers were diagnosed in 11 patients, and no cancers were diagnosed in the surgical biopsy group. There was a nonsignificant linear increase in the likelihood of finding an incidental malignancy with increasing breast density. Thus, these 3626 screening ultrasound examinations yielded 279 unsuspected masses (7.7%) and 11 unsuspected cancers (0.3%). These cancers did not differ significantly in either tumor size or disease stage, from nonpalpable cancers detected on screening mammography in the remaining 7594 patients. The mean tumor size of the cancers detected by ultrasound was 11.9 mm compared with 10.7 mm for those detected mammographically. Eighty-nine percent of ultrasound-detected cancers and 91% of mammographically detected cancers were stage 0 or 1 at the time of diagnosis. In the total population of 11,220 patients, screening ultrasound increased the number of cancers diagnosed from 63 to 74 (+17%) and the number of cancers detected only by imaging from 30 to 41 (+37%).
A cancer yield of 3 per 1000 is not trivial and, at the very least, suggests that additional, larger scale studies of screening with modern ultrasound need to be performed. There is a need for standardization of scanning technique, because ultrasound is highly operator dependent. Care must be taken to include all of the breast tissue and the occasional small subtle lesion must be promptly recognized and characterized. For this reason, many have stated that a radiologist, rather than a technologist should perform the scan. This suggestion raises the issue of the time required to perform these procedures. Kolb et al reported a mean time of 3 minutes and 59 seconds (range: 1 minute 28 seconds to 9 minutes 46 seconds) to perform a bilateral screening ultrasound examination;
Because of its inability to image microcalcifications consistently, ultrasound is unlikely to be useful for screening the population at large. However, as the studies discussed show, there is some promise that, at least in patients with radiographically dense breasts in whom mammography is most limited, ultrasound may have some utility as an adjunctive screening tool. These studies are only preliminary; further studies are needed before any change in current screening practices can be considered.
There is significant overlap in the morphological features of benign and malignant solid masses at ultrasound imaging. For this reason, ultrasound has traditionally not been used to make benign versus malignant distinctions on a routine basis. Palpable solid masses are almost always biopsied (often by fine-needle aspiration) unless imaging reveals an unequivocally benign lesion such as an involuted fibroadenoma. The management of nonpalpable solid masses typically is based on their mammographic appearance, especially the status of the margins. Unless the margins can be confidently characterized as circumscribed, these masses are typically biopsied.
Stavros et al reported on a series of 750 (278 palpable) solid masses prospectively classified as benign, indeterminate, or malignant based on ultrasound findings. [49] The ultrasound features associated with each assessment category were drawn from published studies as well as from the investigators' own experience in a prior, unpublished, retrospective study. A mass was considered benign if it displayed no malignant features and displayed one of three combinations of benign features. If none of these combinations of benign features was present and there were no features of malignancy, it was classified as indeterminate. If at least one malignant feature was present, the mass was considered malignant. A meticulous ultrasound scanning technique was employed, including routine imaging in radial (analogous to spokes of a wheel, with the nipple being the hub) and antiradial (perpendicular to radial) planes to evaluate ductal extension of the mass. All masses were biopsied using an imaging-guided, large-gauge core technique or excision.
At biopsy, there were 125 malignant masses, of which 44 were palpable, and 625 benign masses. ultrasound classified 424 benign lesions as benign (true negatives), 201 benign masses as indeterminate or malignant (false positives), 123 malignant lesions as indeterminate or malignant (true positives) and 2 malignant lesions as benign (false negatives). One of these false negatives was a metastatic deposit from a primary lung carcinoma. Thus, this ultrasound classification scheme had a sensitivity for malignancy of 98.4%, a specificity of 67.8%, a positive predictive value 38.0%, and a negative predictive value of 99.5%. The
During the past 15 years, several important advances in MR imaging technology have combined to increase significantly the applicability of this technique to breast imaging. These advances include the introduction of intravenous, gadolinium-based contrast agents, the development of dedicated breast surface coils, and the elaboration of novel imaging sequences.
Breast MRI is typically performed with a dedicated coil that allows simultaneous imaging of both breasts and the use of a gadolinium-based intravenous contrast agent. Beyond this, there is no standard imaging protocol. Imaging sequences vary widely from institution to institution, and many investigators have their favored sequence. [16] [3] Even the dose and timing of contrast agent administration vary. [16] [3] This lack of standardization makes it difficult to compare the results of published studies, because these variables can affect the success of the technique.
The detection of breast cancer by MRI is possible primarily because the tumor image usually enhances following administration of intravenous contrast. This enhancement is due to the presence of neovascularity induced by the growing tumor. However, interpretation is complicated because images of many benign lesions, and even normal breast tissue, depending on the phase of the menstrual cycle, may enhance, [30] [38] giving rise to false-positive interpretations. Time-intensity curves that describe the enhancement dynamics of a mass on breast MRI, [12] [42] patterns of enhancement (central versus peripheral), [41] [40] and assessment of morphologic
In up to 35% of patients with breast cancer, MRI will demonstrate additional foci of disease not detected on mammography or clinical examination, in some cases leading to a change in therapy. [44] The clinical significance of these unsuspected foci is unknown at this time. Magnetic resonance imaging also depicts the true size of a malignant mass more accurately than mammography. [39] This information, too, may affect therapeutic decision making. Other potential roles for MRI are follow-up of patients after breast conservation therapy [36] or after primary chemotherapy. [1] Also, MRI has been shown to be helpful in identifying a primary breast carcinoma in patients presenting with metastatic axillary adenopathy but no suspicious findings on mammography or on clinical breast examination. [35]
Progress in two critical areas may increase the utilization of breast MRI in
the future. [58]
The first is MRI-guided biopsy techniques. Devices to perform needle biopsy or
needle localization for surgical excision have been developed at a few centers
[31]
[43]
; however, these devices are not widely available. Unless suspicious areas
depicted only on MRI can be readily biopsied, MRI-guided biopsy will have
limited clinical usefulness. The other area of potential usefulness is the
development of small, low-cost dedicated breast MRI units. One such unit is on
the market (Advanced mammography Systems,
Wilmington, MA), and others may follow. Such low-cost units are necessary,
because a major criticism of MRI is its high cost, compared with other breast
imaging modalities.
Film-screen mammography has proven efficacy in breast cancer screening. It does, however, have several distinct limitations. These limitations have spurred research in alternative imaging techniques, each of which seeks to ameliorate one or more of these weaknesses. As discussed in this article, there are many successes to report. However, many of these new technologies and techniques have limitations of their own that must be overcome before their full potential can realized. Given the fast pace of breast cancer-related research, it is fully expected that these challenges will be met in the near future. In this future breast imaging will continue to hold a central role. Screening and lesion characterization techniques will be improved, and imaging-directed biopsy and imaging-directed therapy capabilities will be further refined.
Назад к Русскому Медицинскому Серверу
К медицинскому сайту Участкового