anatomy. MRI differentiates a tumorous lesion from surrounding tissue and permits precise determination of its size. T2-weighted images depict the organs of the true pelvis, and their zonal anatomy, which is the basis for identification of intraand extracervical tumor extension. Furthermore, the calculated apparent diffusion coefficients of cervical cancer have shown to be lower than those of normal cervical stroma, providing increased contrast between the normal cervical stroma and cervical tumor (Hoogendam et al. 2010). In fact, the addition of diffusionweighted imaging improves interobserver agreement, especially when T2-weighted images are equivocal (Chen et al. 2010). MRI is superior to clinical evaluation in assessing tumor size; its measurements are within 0.5 cm of the surgical size in 70–94% of cases (Hricak et al. 1988; Mitchell et al. 2006). For the assessment of small-volume disease, the use of an endovaginal coil has been reported to be more sensitive than an external-array coil (Downey et al. 2016). Studies in the literature report accuracy of 88–97% for MRI in the detection of parametrial invasion as compared to 72% for CT (Brodman et al. 1990; Sala et al. 2007; Scheidler and Heuck 2002). The superior depiction of the vaginal anatomy by MRI with differentiation of the mucosal and muscular layers of the wall results in earlier detection of vaginal involvement, with an accuracy of 90% as opposed to 77–82% for CT (Togashi et al. 1989). Infiltration of the bladder and rectum is demonstrated directly by MRI and identified with an accuracy of 96–100% (Preidler et al. 1996; Ebner et al. 1994). The overall staging accuracy of MRI ranges from 75 to 96% (Sala et al. 2007; Scheidler and Heuck 2002; Ascher et al. 2001).

With regard to lymph node staging, MRI and CT have similar sensitivities of only up to 70%, while specificity is high at approximately 95% (Hricak et al. 1988; Togashi et al. 1989; Boss et al. 2000). Both imaging modalities primarily rely on morphometric criteria for identifying metastatic nodes and therefore fail to detect micrometastases that do not affect lymph node size and shape. Preliminary studies evaluating the use of diffusion-weighted imaging for lymph node assessment have revealed that nodal metastases show significantly decreased apparent dif-

fusion coefficients compared with benign lymph nodes, and abnormal nodes as small as 5 mm may be detected with diffusion imaging (Lin et al. 2008; Kim et al. 2008).

MRI is also a valuable imaging technique for excluding or demonstrating local tumor recurrence. In particular, MRI enables differentiation of postoperative or radiation-induced scars from recurrent tumor. Additional contrast-enhanced dynamic Tl-weighted studies are helpful to differentiate therapy-related changes from tumor tissue. In contrast, CT demonstrates most local recurrences only when they produce a mass effect or infiltrate adjacent structures or organs (Heron et al. 1988).

2.2\ Imaging Technique

2.2.1\ MRI

Streamlined technical guidelines for uterine cervical cancer staging and follow-up with MRI were recently issued by the female imaging subcommittee of the ESUR (European Society of Urogenital Radiology) and may serve as an excellent guide to achieve state of the art examinations (Balleyguier et al. 2011).

A brief gynecologic history should be obtained prior to the MRI examination. As the morphologic appearance of the uterus varies with the patient’s hormonal status, information on the phase of the menstrual cycle or postmenopausal status as well as on hormone therapy should be gathered. Moreover, the history should comprise information on pregnancies and cesarean sections as well as on invasive diagnostic procedures such as cervical conization or curettage. In patients undergoing followup MRI, information on earlier pelvic surgery or radiochemotherapy is important. The radiologist needs these data to correctly interpret the morphologic MR appearance.

MRI is performed with the patient in the supine position. Fasting is not necessary prior to the examination.

Moderate bladder filling will straighten an anteflexed uterus. Too much bladder filling may lead to restlessness during the course of the examination and may even make it necessary to discontinue the examination.

For an optimal image quality, artifacts caused by intestinal peristalsis should be minimized, in general by administration of a spasmolytic agent at the beginning of the examination. The usual agent is butylscopolamine bromide (Buscopan) administered intravenously or intramuscularly at a dose of 40 mg. This spasmolytic agent has an elimination half-life of 2–3 h. The onset of spasmolytic activity is immediately after IV administration while it is delayed by a few minutes after IM administration. Alternatively, patients with contraindications to butylscopolamine bromide (hypersensitivity, glaucoma, driving immediately after the examination) are given an IV dose of 2 mg glucagon (GlucaGen). Technically, motion artifacts can be reduced by rapid image acquisition.

A high signal-to-noise ratio (SNR) and a high spatial resolution are important for optimal pelvic evaluation by MRI. For this reason, body phased-array surface coils are most frequently used. The resolution can also be improved with the use of a small field of view (FOV), for instance 20 × 20 cm, in combination with phase oversampling to prevent wrap-around artifacts (aliasing). With surface coils being highly susceptible to artifacts caused by respiratory motion of the abdominal wall, all sequences must be acquired with presaturation of the abdominal wall.

A valuable alternative to the phased-array surface coils, in case of small-volume cervical carcinomas, is the endovaginal MRI technique consisting of a ring-design endovaginal receiver coil of solenoid geometry inserted into the vagina and positioned around the cervix (deSouza et al. 2006).

The imaging area should comprise not only the pelvis but also the abdomen up to the renal hilum in order to include the para-aortic lymph nodes. This applies especially to patients with cervical cancer stage IIB and above.

The MRI examination begins with a localizer scan in transverse, sagittal, and coronal orientation, followed by T2-weighted imaging in two planes. T2-weighted sequences have the highest soft-tissue contrast and thus provide most of the information on the localization and extent of a

cervical carcinoma. They are the basis of any pelvic MRI examination.

The first T2-weighted sequence should be acquired in the sagittal plane and cover the uterus and vagina to the pelvic floor. This sequence should be acquired with high resolution using thin slices and a small FOV, i.e., a 384 × 256 matrix size, a FOV of 240 mm, and a slice thickness of 4–5 mm.

The sagittal T2-weighted images may serve to plan the transverse angulated T2-weighted sequence. The transverse sequence should be angulated for alignment perpendicular to the axis of the cervical canal. As with the sagittal sequence, the imaging field in transverse orientation extends from the fundus uteri to the pelvic floor. Images should be acquired with a slice thickness of 4–5 mm, a 384 × 256 matrix, and a phase resolution of at least 75% (Table 4 and Fig. 11).

The cervix uteri is normally anteversed and forms an angle of about 90° with the axis of the vagina. The uterine corpus is flexed forward, resulting in an angle of 70°–100° relative to the cervix. The degree of anteflexion varies with bladder filling and is also affected by the size of the uterine corpus. From puberty onwards, the cervix-to-corpus ratio is 1:2, and the corpus becomes again smaller after menopause and descends into the true pelvis. The positions of the uterus are shown in Fig. 6.

Angulated image acquisition ensures optimal depiction of the cervix and parametria and their topographic relationships (Fig. 7). It is important that the angulation does not exceed 45° to avoid acquisition in coronal orientation with reversal of left and right.

In cases of vaginal involvement with the risk of parametrial infiltration through the paravaginal tissue from below, additional angulation perpendicular to the vagina is useful. For optimal evaluation of vaginal infiltration, the vagina may be distended with ultrasound gel. Involvement of the pelvic floor muscles in advanced tumors is evaluated on coronal T2-weighted images, which is especially suited for evaluation of the levator ani muscle. Information on muscle involvement is important for planning of the surgical procedure.

For evaluation of the pelvic sidewall and lymph node staging, an additional proton-density or a T1-weighted sequence in transverse orientation should be performed (see Fig. 10). The acquisition starts at the level of the aortic bifurcation and extends to below the pelvic floor. A slice thickness of 6 mm is used, with a 512 matrix and a phase resolution of at least 60%. Complete coverage of the inguinal lymph nodes should be attempted in patients with cervical cancers involving the lower third of the vagina (stage

IIIA or higher), which are more likely to be associated with inguinal lymph node metastasis.

Although intravenous contrast medium administration is not routinely used for primary staging of cervical cancer, recent studies have shown that contrast-enhanced T1-weighted imaging may be superior to T2-weighted imaging in terms of cervical cancer tumor localization and tumor margin detection, especially for patients with small tumors (Akita et al. 2011) (Figs. 8 and 9).

Fig. 7  Angulation. T2w TSE images in sagittal orientation of a normal cervix. The cervix is shown in strictly axial orientation (a) and in angulated axial orientation (b). With angulated image acquisition, a superior anatomic

view is obtained of the structures that are crucial for staging (cancer, cervical stroma, parametria) and the cervix is depicted in more slices

Contrast-enhanced images can also improve the diagnostic accuracy of identifying tumorous infiltration of the urinary bladder or rectum. The extent of infiltration is visualized as a disruption of the muscular wall of these organs, which is of lower signal intensity on T1-weighted images. It has been shown that contrast-enhanced MRI can improve the differentiation of an edematous stromal reaction of the vesical or rectal wall (no enhancement) from tumorous infiltration (positive enhancement) (Hawighorst et al. 1997). Contrastenhanced pelvic MRI is performed by acquiring T1-weighted sequences before and after contrast medium (CM) administration. The CM generally used is an unspecific gadoliniumbased low-molecular agent administered at a dose of 0.5 mmol/kg body weight. The con- trast-enhanced T1-weighted study can be planned by adopting the imaging area and orientation of the T2-weighted sequences.

CM administration is indicated to differentiate recurrent tumor from postoperative and postactinic changes at follow-up MRI (Fig. 10). A dynamic T1-weighted postcontrast study with

repeated acquisitions enables temporally resolved quantification of CM enhancement.

Diffusion-weighted imaging (DWI) is a functional MR technique that helps characterization of biological tissues based on water proton mobility, providing valuable information on extracellular space tortuosity, tissue cellularity, and cellular membrane integrity (Bammer 2003; Liu et al. 2015). The architectural malformations associated with cancer result in shrinkage of the extracellular space, which in turn restricts the diffusion of water. The calculated diffusion of water in vivo is expressed as apparent diffusion coefficient (ADC). In the last decade DWI has gained great acceptance and has nowadays become part of the routine imaging protocol for gynecologic oncology imaging. Several studies have demonstrated the value of DWI in the detection, staging, and characterization of malignant tumors of the uterine cervix. DWI enables accurate discrimination of cervical cancer from benign cervical lesions and normal cervical tissue, which may be useful in patients with isointense tumors in T2-weighted images or in patients with small-

volume cervical cancer. Recent studies have shown the potential ability of DWI to indicate the histologic type of cervical cancer as well as the grade of tumor differentiation (Liu et al. 2013, 2015; Thoeny et al. 2012).

Various other sequences are available to answer specific queries and provide specific

information as required according to tumor stage. Cervical cancer extending beyond the cervix and recurrent tumors are associated with a significantly higher risk of metastatic spread to paraaortic lymph nodes. These lymph nodes are best evaluated with high-resolution respiratory-trig- gered T2-weighted sequences for imaging of the