5.2.1 Acquisition Plane

Selection of the primary imaging acquisition plane is critical when choosing a breast DCE-MRI protocol and may be the only acquired imaging plane needed. High-resolution multiplanar reformatted images can be reconstructed from the acquired plane, eliminating the need for additional scans and shortening the scanning time. Axial and sagittal acquisition planes are generally preferred to the coronal plane, given that the anatomic segments of the breast are oriented from the nipple to the chest wall and are not aligned segmentally in the coronal plane. Selection of the acquisition plane is particularly important when assessing DCIS, because axial and sagittal imaging may demonstrate linear or segmental enhancement to advantage, whereas coronal imaging may only show cross-section of ducts. Selection of the field-of-view (FOV) depends on the primary acquisition plane and in general should be chosen to be as small as possible, in order to maximize in-plane spatial resolution while also including all breast tissue and axillary regions. Axial acquisition is the most commonly used primary plane in the United States, providing easy bilateral side-by-side assessment of enhancement at identical postcontrast injection time points.

5.2.2 Imaging Protocol

The current standard-of-care breast DCE-MRI protocol consists of a precontrast T2w acquisition followed by a dynamic sequence encompassing a series of T1w acquisitions obtained before and for approximately 5 to 7 minutes after the intravenous administration of a GBCA. In most practices, temporal resolution has remained between 60 and 120 seconds, with a maximum temporal resolution of 240 seconds allowed in the Breast Magnetic Resonance Imaging Accreditation Program. ¹ Standard diagnostic imaging protocols capture diffusion rather than perfusion of contrast material into enhancing lesions. The ACR Accreditation Program requires both a T2w (bright fluid) series and a multiphase T1w dynamic series. The dynamic series must include a precontrast T1w series, an early-phase (first) postcontrast T1w series to be completed within 4 minutes of contrast injection, and a late-phase T1w series with features matching the precontrast T1w series.

Multiphase T1-Weighted Sequence (DCE-MRI)

This key sequence is used for lesion detection, characterization, and evaluation of lesion contrast enhancement over time. Identical scan parameters must be used for the pre- and postcontrast multiphase T1w series sequences, so that image registration can be performed. A 3D gradient echo (GRE) pulse sequence with a short time of repetition (TR) is commonly used for multiphase T1w imaging. It is also recommended that in order to avoid any confounding T2 contrast, the GRE pulse sequence should be spoiled. ² Precontrast images are subtracted from postcontrast images and signal differences between the sequences can be directly compared. Most practices in the United States use fat suppression in the T1w dynamic series, and subtraction images are helpful for highlighting signal from contrast agents. If the dynamic series is performed without fat suppression, then subtraction images are essential for elimination of the fat signal. Knowledge of the k-space sampling pattern is helpful, most Cartesian sequences use rectilinear sampling with the center of the acquisition capturing the high-frequency data providing optimal contrast resolution. When setting up a breast MRI protocol, there is no recommended consensus as to the number of postcontrast acquisitions within a DCE series, or requirements for the total length of the acquisition time. A minimum of at least two postcontrast sequences are needed to provide a basic assessment of kinetic features, and these should include measures of the initial slope of enhancement and the delayed phase.

Measures of the rate of uptake and washout of contrast media within breast lesions contain diagnostically useful information, and microvascular density is a key factor in determining the initial rate of contrast media uptake and the heterogeneity of enhancement within tumors. ³ The shape of a signal intensity versus time curve (TIC) (or signal time course or kinetic curve) has shown to be useful in the classification of enhancing lesions. ⁴ Lesions with a fast uptake of contrast followed by a washout phase are more likely to be malignant, while lesions with a slow but persistent uptake are more likely benign. Discrete thresholds are used to classify two parts of the kinetic curve: the initial uptake and the delayed phase (Fig. 5‑1).

5.3.1 Deposition of Gadolinium in Tissue

It is well known that not all gadolinium chelates are entirely stable in vivo. There are two structurally distinct categories of commercially available GBCAs: linear and macrocyclic. The structure of a linear GBCA provides a polyamino-carboxyl acid backbone wrapping around the gadolinium ion which is incompletely enclosed, whereas in the macrocyclic structure, the gadolinium ion is “caged” in the ligand. Although patients with normal renal function clear almost all gadolinium from the circulation within a few hours by renal excretion, some free gadolinium may disperse into tissues. Intracranial gadolinium retention has been reported recently showing that trace amounts of gadolinium accumulation, as evidenced by concomitant T1 shortening on MRI, are present in the brains of patients who have received multiple doses of GBCA. ^{7 , 8} This effect is seen in patients without renal or hepatic dysfunction and is predominantly found in the globus pallidus, thalamus, pons, and dentate nucleus. ⁹ The magnitude of the effect is proportional to the cumulative lifetime patient dose. Some research suggests that this phenomenon occurs more commonly when linear agents with weaker chemical binding to gadolinium are administered, compared to the stronger chemical binding of macrocyclic agents. ¹⁰ Visible intracranial changes on T1w images may be present after only four exposures of a GBCA, and these findings often remain for years after the MR studies took place. In 2015, the U.S. Food and Drug Administration (FDA) instituted a review of gadolinium contrast agents to determine whether a safety risk to patients exists. On March 10, 2017, the Pharmacovigilance Risk Assessment Committee (PRAC) of the European Medicines Agency (EMA) formally submitted its recommendation to suspend use of four linear GBCAs due to the potential risk of gadolinium accumulation within humans. ¹¹ The PRAC noted that no clinical diseases or even symptoms have been reported in connection with gadolinium deposition, but the committee decided on a “precautionary approach,” noting the paucity of data as well as concern regarding gadolinium’s past history of association with nephrogenic systemic fibrosis (NSF) in patients with kidney impairment. A response to the European PRAC recommendations was published by the ACR on April 4, 2017. This report concluded that “after extensive review of the PRAC position and voluminous other materials, the ACR Committee on Drugs and Contrast Media disagrees with the PRAC recommendation.” The ACR report concluded that there is abundant evidence that the amount of gadolinium deposited in tissues after a single GBCA dose is incredibly small and found no compelling evidence that any GBCA (including linear agents) pose a safety risk. The ACR additionally endorsed the need for additional research efforts to be directed toward a greater understanding of the mechanisms, cellular effects, and clinical consequences of gadolinium tissue deposition. ¹² On September 8, 2017, the Medical Imaging Drugs Advisory Committee (MIDAC) of the U.S. FDA voted to recommend the addition of a warning to labels of GBCAs used during MRI. The new labels will warn that gadolinium may be retained in various organs, including the brain; after use of these agents, linear GBCAs exhibit a greater risk than macrocyclic agents. The warning will also note a greater risk in specific patient populations that include the fetus and children and patients who receive large cumulative doses of GBCAs, such as women undergoing annual GBCA-enhanced MRI because of high risk for breast cancer. At the present time however, although there is clear evidence that residual amounts of gadolinium may be retained in the brain and other parts of the body after administration of GBCA for MRI examinations, there is no definitive evidence that residual gadolinium is associated with adverse health effects. Ongoing research on this topic is underway.

5.4.1 Ultrafast Techniques for Diagnostic Use

Protocols, referred to as “ultrafast (UF)” or “accelerated” imaging, have been investigated and developed intermittently over the years and are now once again being used in clinical practice. These protocols are aimed toward image sampling of early kinetics within the first minute postcontrast injection at a fast rate (3–8 seconds per time point) for a bilateral scan, as shown in Chapter 3. Current diagnostic DCE-MRI examinations at UC include a high temporal resolution, lower spatial resolution (UF) acquisition during the first minute following contrast injection, followed by a standard, higher spatial resolution acquisition, achieving a total acquisition time of 7 minutes. Lesion conspicuity in UF images is often greatest within the first 30 seconds after contrast media injection, especially when marked BPE is found on later images. UF methods allow measurement of the rate of lesion enhancement with respect to time of initial arterial enhancement in the breast rather than time of injection, thus reducing dependence on global variables such as cardiac output. UF imaging in the first minute postcontrast injection can be incorporated into a standard-length diagnostic protocol, and delayed-phase kinetics can thus also be measured. ¹⁷ This approach to UF imaging uses standard Cartesian k-space sampling, which differs from “view-sharing” methods, that are often used in other clinical DCE-MRI high- resolution protocols. Keyhole approaches and other “view-sharing” methods such as “TRICKS” and “TWIST” typically involve scanning the center of k-space at relatively high temporal resolution, and the outer portions of k-space at much lower temporal resolution. ^{18 , 19 , 20} Each image in the dynamic series is then produced from data obtained from different parts of k-space, scanned at different times, resulting in high-quality images, but quantitative analysis of the acquired data with these sequences is difficult. Problems arise when data are analyzed to determine which features are enhancing at what strength and at what rate, particularly in the case of small features (e.g., lesions) with sharp and irregular edges. Information obtained from the initial 10 seconds after uptake initiation is combined with information from other time periods, and the diagnostic utility is thus diluted. For example, if a narrow blood vessel, the rim of a cancer, or speculations on a suspicious lesion appear to enhance rapidly, it is difficult to determine the true rate of enhancement. Compressed sensing methods have similar problems and produce artifacts. ^{21 , 22} Further examples of UF DCE-MRI incorporated into a standard diagnostic acquisition are shown (Fig. 5‑6, Fig. 5‑7, Fig. 5‑8, Fig. 5‑9, Fig. 5‑10).

UF DCE-MRI allows new approaches to quantitative data analysis, including measurements of the initial rate of lesion enhancement. Time of arrival of the contrast bolus can be accurately measured relative to the initial enhancement in the aorta. This potentially powerful approach can correct for variations in contrast media injection speed and cardiac output, allowing for more standardized measurements. UF imaging may also have other advantages; high temporal resolution protocols allow propagation of the contrast media bolus into a lesion allowing visualization of feeding arteries and draining veins and providing simplified and more accurate measurements of the arterial input function and K_trans. Preliminary research suggests that there are significant differences between benign and malignant lesions in parameters that describe contrast media uptake kinetics in the first minute postinjection.