Optimization of MR imaging for pretreatment evaluation of patients with endometrial and cervical cancer.

The above described T2 weighted sequences may be supplemented with 3D T2 fast spin echo images. TP [This technique consistently provides images with high SNR and contrast to noise ratio (CNR), excellent T2 contrast and superior anatomic definition, in addition to the ability to retrospectively generate multi-planar images] ( Fig. 15 ). Images are acquired with 2 mm thickness and should be preferably in the same plane as the oblique axial T2 FRFSE images. Occasionally, this technique is limited by the inability to apply the no phase wrap function, and consequently a larger FOV needs to be acquired to prevent aliasing. This limits the resolution of the acquired images.

Metastasis to the regional lymph nodes is one of the most important prognostic factors in endometrial cancer and included in the FIGO surgical staging of endometrial cancer. The presence of nodes upstages endometrial cancer to a stage IIIC1 or IIIC2, depending on whether pelvic or para-aortic nodes are involved ( 3 ). The FIGO staging of cervical cancer is clinical and does not include adenopathy. However, nodal involvement has significant prognostic implications and is very important in treatment planning ( 20 , 28 , 49 ). Assessment of nodal involvement in using cross sectional imaging techniques continues to rely on nodal size, which has significant limitations, with sensitivity in the range of 38-89% and specificity from 78-99% ( 20 , 49 - 53 ). TP [The incorporation of morphologic features of nodal involvement best seen on high resolution T2 WI, including internal heterogeneity, spiculated nodal margins, necrosis, signal intensity comparable to the primary tumor, has improved accuracy of nodal involvement evaluation in rectal cancer and is potentially applicable to endometrial and cervical cancer] ( 54 ). The high resolution coronal T2WI provide the best assessment of pelvic nodes that may be involved in cervical and endometrial cancer i.e. parametrial, obturator, internal, external iliac, and common iliac nodal stations ( Fig. 17 ). Large FOV axial T2 fast spin echo series from the top of the kidneys through L3 (FOV 30-38 cm, slice thickness 5 mm) allows assessment of para-aortic lymphadenopathy as well as presence of hydronephrosis. The main as well as optional MRI sequences discussed in this article are presented in the Table 3 . More precise tailoring of the MRI protocol should be done on the basis of the type of gynecologic malignancy with optional sequences used as needed to improve cancer staging ( 7 , 27 ).

There is no consensus in the literature regarding use of vaginal contrast ( 2 , 9 , 20 , 28 , 48 ). Therefore, the use of vaginal contrast remains optional. Vaginal gel is useful for the evaluation of cervical cancer patients, especially in the subgroup who do not undergo evaluation under anesthesia. About 20-30 ml of warm ultrasound gel is placed in the vagina after positioning the patient on the table. Usually, vaginal contrast is well tolerated and does not cause any significant discomfort. Vaginal opacification with gel provides high signal intensity on T2W images and enables excellent definition of vaginal fornices and cervix, allowing for accurate assessment of vaginal involvement, especially in tumors with an exophytic cervical component ( Fig. 16 ) [CME#10].

Endometrial carcinoma is the fourth most common cancer among women and the most common cancer of the female reproductive tract with an estimated 49,560 new diagnoses and 8,190 deaths in the US in 2013 with average age at the time of diagnosis of 61 years ( 1 ). The increasing incidence observed in recent years, is thought to be due to both higher life expectancy and rising rates of obesity. Cervical carcinoma is the third most common gynecologic malignancy with 12,340 new cases and 4,030 deaths in 2013 and average age at the onset 48 years ( 1 ). The widespread use of screening with the Papanicolaou smear, and effective treatment of carcinoma in situ have led to a significant decline in cervical cancer in the developed world ( 2 ). The FIGO staging of endometrial cancer is surgical, requiring hysterectomy, bilateral salpingo opherectomy, node dissection, peritoneal washing and omental biopsy for appropriate staging ( Table 1 ) ( 3 ) [CME#1]. But distinct from the requirements of FIGO staging is the clinical management of patients with endometrial cancer. Patients that present with an early stage low or intermediate risk disease, which includes stage IA grade 1, 2 and 3, and stage IB grade 1 or 2 endometroid histology can be treated appropriately with minimally invasive laparoscopic hysterectomy and bilateral salpingoophorectomy. This approach leads to reduced morbidity and hospital stay, and in this group has comparable outcomes to the more extensive surgical resection, which should be reserved for high risk patients with stage IB grade 3 endometrioid, or stage II and above, and all grades non-endometrioid histology, or ( 4 - 7 ). However, the effective implementation of this treatment approach relies heavily on accurate pre-surgical staging. MRI, particularly using the multi-parametric approach has been shown to be reliable in terms of assessment of the key treatment determinants, which are depth of myometrial invasion and cervical stromal involvement ( 7 - 11 ) [CME#2]. Although not part of FIGO, staging MRI is recommended by the National Cancer Institute of France, the European Society of Radiology Guidelines and the Royal College of Radiologists. ( 12 - 14 ). Assessment of lymph node involvement in endometrial cancer based on size criteria has significant limitations. However, the incidence of nodal involvement correlates with depth of myometrial invasion and cervical stromal involvement, and consequently they can be used as surrogates for determining the need for lymph node dissection ( 13 , 15 , 16 ). Lymph node metastasis increases from 3% when the depth of tumor invasion is less than 50% of the myometrial thickness to 46% with deeper involvement ( 5 , 17 , 18 ). The role of imaging in endometrial cancer has potentially received a further boost, by the recent modification of the FIGO staging in 2009 ( 3 ). The new staging system has combined superficial (50% thickness) are now termed stage IB. In addition the definition of stage II has changed with removal of cervical mucosal involvement as a determinate of upstaging to now only cervical stromal invasion used to define stage II tumors. Distinguishing disease confined to the endometrial cavity from superficial myometrial invasion and defining cervical mucosal involvement, part of prior FIGO staging, was a limitation of imaging. Elimination of these categories could potentially improve the staging accuracy of MR imaging in endometrial cancer ( 19 ). Cervical cancer is the single gynecologic malignancy still staged clinically according to the revised 2009 FIGO classification ( Table 2 ) ( 20 ) [CME#3]. However the committee encourages the use of imaging, if available, as clinical staging is inaccurate in 22-75% ( 21 ). The use of MR imaging enables a more appropriate triaging of patients to hysterectomy if the tumor is confined to the cervix and < 4 cm in size, or chemo-radiation if tumor size exceeds 4 cm or parametrial invasion is present. Although recent multi-institutional trials have raised concerns about the accuracy of cross-sectional imaging in the staging of early cervical cancer (stage < IIB), MRI still remains the best imaging technique for assessment of tumor size, with a high negative predictive value in excluding parametrial invasion ( 22 - 26 ). The realization of these objectives however is dependent on the use of appropriate MRI technique. Recent publications underscore the value of multiparametric MR imaging combining sagittal and oblique axial T2 weighted images (T2WI), DCE and DWI in staging and treatment stratification of patients with gynecologic malignancies ( 7 , 27 ). In this article we review optimized MR protocols incorporating high resolution T2WIwith an emphasis on the importance of good quality multi-planar images, DCE-MRI) and DWI.

Optimization of the MR imaging protocol with use of thin section high resolution multi-planar T2 weighted images, addition of simple modifications such as double oblique T2 weighted images, supplemented by diffusion weighted imaging and dynamic contrast enhanced MRI improves staging and treatment planning of the endometrial and cervical cancer.

The combination of T2 WI and DCE-MRI offers high accuracy in staging endometrial cancer in the range of 83-91% ( 2 , 9 , 28 , 29 ), with only a few dissenting papers reporting no added benefit with post contrast images ( 30 , 31 ). More recent studies have found oblique axial fused T2 and DW images have a high accuracy in assessing depth of myometrial invasion, with some papers reporting not only superior accuracy to DCE-MRI, but also higher inter-observer agreement ( 32 - 34 ). The incorporation of all three sequences may represent the most comprehensive approach to the preoperative staging of endometrial cancer ( 7 , 27 ). A significant component of the reliability of the multiparametric approach is the acquisition of good quality multi-planar images, most particularly two planes orthogonal to the tumor obtained, if possible with each sequence, but definitely with the T2 weighted and post contrast T1 weighted sequences. In addition the orthogonal T2WI, DCE-MRI and DW images should co-register by slice location so as to enable correlation of findings on the different sequences. This improves staging accuracy ( 4 ). TP [T2WI is the key sequence in the evaluation of myometrial invasion, since this sequence provides depiction of the uterine zonal anatomy with the intermediate signal tumor well delineated against the low signal intensity junctional zone] ( Fig. 1 ) ( 7 - 9 , 11 ) [CME#4]. However T2WI may be limited in post-menopausal patients, where the zonal anatomy of the uterus is less well defined or if the tumor is isointense to the myometrium. The use of thin section (3 mm) oblique axial and sagittal T2WI (FOV 20-22 cm) is well established in the staging of endometrial cancer ( 27 ). Our suggested modification to the imaging protocol is to obtain high resolution T2 weighted fast relaxation fast spin echo (FRFSE) images in three planes, sagittal, coronal and oblique axial. In addition, since the position of uterus is notoriously variable, an oblique axial image based only on the sagittal images occasionally cannot provide an orthogonal view to the tumor. An additional sequence maybe useful, for instances when the uterus is tilted to the left or right of the midline. In such cases, T2 oblique axial images angled off both the sagittal and coronal planes create a “true oblique axial” that is correctly positioned along the true axis of the uterus. This is a “double oblique” sequence as it is oblique in two planes, the sagittal and coronal ( Fig. 2 ). TP [HR double oblique images allow a true orthogonal view of the uterus, with a potential to avoid volume averaging, and improve assessment of myometrial invasion] ( Fig 3 ). To ensure good spatial resolution and signal to noise ratio (SNR) the images should be acquired using a surface coil appropriately centered over the uterus, using a 20-24 cm field of view (FOV) with 3 mm contiguous cuts. The FOV should be adjusted to ensure appropriate SNR, and if needed, it should be increased as this single step can double the SNR ( Fig. 4 ). It is also valuable to have the patient fast 4-6 hours prior to the scan and empty the bladder before going onto the MR scanner to reduce motion. Antiperistaltic agents, such as hyoscine butylbromide or glucagon, are used in many centers to reduce motion artifacts from bowel peristalsis ( 27 ). The phase and frequency direction can also be adjusted to avoid motion artifacts from bowel loops or the bladder wall. A wide anterior saturation fat suppression pulse can generally eliminate motion artifacts from the anterior abdominal wall ( Fig. 5 ). TP [The advantage of using multi-planar high-resolution imaging is a greater confidence in assessment of tumor stage by improved spatial resolution and the ability to confirm the extent of disease in more than one plane which is essential to accurate staging] ( Fig. 6 ) ( 4 , 33 ) [CME#5]. Although T2 weighted images are essential, they often prove inadequate due to poor tumor-myometrium contrast, poor definition of the junctional zone particularly in postmenopausal patients, adenomyosis and leiomyomas that compromise accurate staging ( Fig. 7 ). DCE-MRI and DWI can occasionally help to overcome these potential pitfalls. The relative merits of T2WIand DCE-MRI appear to be related the menopausal status, with T2-weighted scans showing a greater accuracy in staging in premenopausal patients and DCE-MRI in postmenopausal patients ( 35 ). With the exception of a few dissenting reports it has been widely accepted that the use of DCE-MRI improves accuracy of tumor staging in endometrial cancer. This is essentially a consequence of the improved tumor-myometrial contrast generally seen on the delayed 2-4 minute scans, where most endometrial tumors appear hypointense against the enhancing myometrium ( Fig. 8 ). There are additional benefits of the post contrast scans, small tumors that may be difficult to define on the T2WI may appear hypervascular on the early arterial phase images and in patients with loss of the junctional zone or adenomyosis. These post contrast images can assist in assessment of depth of myometrial invasion, while the definition of the intact enhancing cervical mucosa excludes cervical stromal invasion ( 27 , 33 ). A limitation encountered with DCE-MRI images is that some tumors may be isointense relative to the myometrium on the equilibrium phase (2 min post injection), negating the benefit of this sequence. The efficacy of DCE-MRI in staging relies on obtaining images in two orthogonal planes ( 33 ). This is achieved by acquisition of images in the sagittal and oblique axial plane. Most commonly dynamic fat suppressed 3D T1 fast spoiled gradient echo images (3D FSPGR) are acquired in the sagittal plane. These may be obtained at 30, 60, 120 seconds or by scanning continuously through the uterus for 2 minutes. This is followed by delayed (3-4 min) oblique axial fat suppressed 3D T1 weighted images along the axis of the uterus preferably with the same slice positioning as the oblique axial T2 weighted images ( Fig. 9 ). DWI is a functional imaging technique whose contrast derives from the differences in restriction of motion of water molecules. The clinical utilization of DWI in gynecologic malignancies has been steadily increasing over the recent years ( 36 - 39 ). Studies have shown a significantly lower ADC value for endometrial cancer (0.86-0.98 × 10×-3/mm2/s) then normal endometrium (1.28-1.65 × 10×-3/mm2/s) and higher grade endometrial cancers exhibit a tendency toward lower ADC values than more well differentiated tumors ( 36 , 37 , 40 , 41 ). These distinct ADC values of endometrial malignancies can help in localizing the tumor in the midst of the normal endometrium, and this facilitates tumor staging. Recent reports evaluating the efficacy DWI in staging concluded that DWI with relatively high b value (1000 sec/mm2) fused with T2WI provided accurate assessment of myometrial invasion, which reportedly improves the accuracy of T2WI and DCE-MRI ( 32 ). The impact of this technique was particularly evident with tumors that are iso-intense to the myometrium on DCE-MR images or where intravenous contrast cannot be utilized. DW images should be obtained with variable b values in the range of 50 and 500-1000 preferred in the pelvis. The images should ideally be acquired in the same plane and with a comparable FOV as the oblique axial T2 weighted and DCE images and then fused. If the images cannot be fused, the slice locations should be co-registered on all three sequences to permit correlation ( Fig. 10 ). During interpretation, it is important that DW images always be read in conjunction with apparent diffusion coefficient (ADC) maps to avoid misinterpretation related to T2 shine through.