OBJECTIVE. In this article, we review the role of MRI in the imaging of malignant neoplasms of the uterine corpus and cervix, describing its role in staging, treatment planning, and follow-up.
CONCLUSION. MRI is not officially incorporated in the International Federation of Gynecology and Obstetrics (FIGO) staging system, but is already widely accepted as the most reliable imaging technique for the diagnosis, staging, treatment planning, and follow-up of both endometrial and cervical cancer. MRI protocols need to be optimized to obtain the best results and avoid pitfalls.
The role of MRI in gynecologic oncology has evolved during the past two decades. There is now a substantial body of evidence that MRI is useful in evaluating malignant conditions of the pelvis [1, 2]. MRI has been shown to be superior to CT in staging of endometrial and cervical carcinoma. In addition, there is evidence that MRI may aid in differentiating radiation fibrosis from recurrent tumor . The accuracy of MRI assessment of lymph nodes is similar to that of CT; both rely on size criteria to detect the presence of metastases . However, more recently, lymph node–specific contrast agents have emerged as useful tools for determining the presence of metastases in the lymph nodes .
MRI has been shown to minimize costs in some clinical settings by limiting or eliminating the need for further expensive or more invasive diagnostic or surgical procedures [6, 7]. In this article, we review the role of MRI in staging, treatment planning, and follow-up of malignant neoplasms of the uterine corpus and cervix.
Malignant Neoplasms of the Uterine Corpus
Adenocarcinomas arise from the uterine epithelium and constitute 90% of endometrial cancers. The remaining histologic types of endometrial carcinoma include adenocarcinoma with squamous differentiation, adenosquamous carcinoma, clear cell carcinoma, and papillary serous carcinoma. Uterine sarcomas are rare tumors of mesenchymal origin accounting for 2–6% of all uterine malignant tumors . The most common histologic variants are endometrial stromal sarcoma, mixed müllerian tumors, and leiomyosarcoma. Primary uterine lymphoma is very rare, occurring in only 1% of patients with lymphoma. Metastases to the uterus from nongynecologic neoplasms are rare, with breast and the gastrointestinal tract being the two most common primary sites.
Endometrial carcinoma is the fourth most common female cancer and the most common malignancy of the female reproductive tract . In 2007, 39,080 new cases and 7,400 deaths are expected in the United States . The incidence is rising because of increased life expectancy and obesity. Five-year survival rates vary between 96% for stage I disease and 25% for stage IV disease . The prognosis of women with endometrial carcinoma depends on a number of factors, including stage, depth of myometrial invasion, lymphovascular invasion, nodal status, and histologic grade . Preoperative evaluation of these prognostic factors helps in subspecialist treatment planning .
Imaging technique and patient preparation are important to obtain optimal results. Patients are usually instructed to fast for 4–6 hours before the MRI examination to limit artifact due to small-bowel peristalsis. An anti-peristaltic agent (hyoscine butyl bromide or glucagons) may be administered to the patient before imaging as an alternative to fasting. Ideally, the patient is asked to empty the bladder before going on the MR scanner. A full bladder may degrade T2-weighted images because of ghosting and motion artifacts. Patients are imaged in the supine position using a pelvic surface array multichannel coil.
The basic MRI protocol (Table 1) includes axial T1-weighted spin-echo images with a large field of view to evaluate the entire pelvis and upper abdomen for lymphadenopathy and bone marrow changes; high-resolution T2-weighted fast spin-echo (FSE) images in the axial and sagittal planes for the evaluation of the primary tumor; and dynamic contrast-enhanced T1-weighted images (small field of view) in the sagittal and axial oblique planes to evaluate the extent of myometrial and cervical involvement.
TABLE 1: Pearls and Pitfalls of MRI for Staging of Carcinoma of the Endometrium and Cervix
Dynamic multiphase contrast-enhanced 3D T1-weighted imaging is more accurate than T2-weighted imaging for the assessment of the depth of myometrial invasion
Loss of junctional zone definition; band of subendometrial enhancement seen in only 50-60% of cases
Early phases better to visualize subendometrial enhancement band (stage IA vs IB)
Distension of endometrial cavity by polypoid tumor compressing myometrium
Maximum inner—outer myometrium contrast at 1 min (stage IB vs IC)
Poor tumor—myometrium contrast
Maximum tumor—myometrium contrast at 2-3 min (stage IB vs IC)
Tumor extending to the uterine cornu
Enhancement of cervical mucosa on delayed images (4-5 min) excludes cervical stroma invasion
Presence of microscopic disease
Second imaging plane necessary for accurate evaluation of depth of myometrial invasion
Accurate estimation of tumor size by MRI (within 0.5 cm of measurement at pathology)
Parametria are located lateral and only lateral to the cervix
Intact low-signal-intensity stromal ring excludes parametrial invasion
Loss of low-signal-intensity stromal ring indicates full stromal but not parametrial invasion
Recurrent vaginal vault tumor has the same signal intensity as the primary tumor
Overestimation of parametrial invasion on T2-weighted images due to postbiopsy hemorrhage or with large tumors due to stromal edema
Reconstitution of normal cervical anatomy and low-signal-intensity cervical stroma indicate complete response to radio- or chemotherapy
Early radiation change (within 6 months) and presence of infection may show enhancement
Dynamic contrast-enhanced MRI improves detection of small tumors and helps in differentiating tumor recurrence from radiation fibrosis
High-resolution T2-weighted FSE sequences perpendicular to the long axis of the uterine corpus are favored for the evaluation of primary tumor and myometrial invasion . Sagittal and oblique axial multiphase IV contrast–enhanced 3D T1-weighted fat-saturated sequences through the uterine corpus are routinely used to improve staging accuracy. The early enhancement phases (0 and 1 minute) allow identification of the subendometrial zone, which enhances earlier than the bulk of the myometrium and corresponds to the inner junctional zone. Identification of this zone is especially important in detecting early myometrial invasion because the junctional zone often becomes indistinct in post-menopausal women . The equilibrium phase (2–3 minutes after injection) allows better evaluation of deep myometrial invasion , whereas the delayed phase (4–5 minutes) enables better evaluation of cervical stroma invasion . The tumor–myometrium interface should be assessed in at least two planes.
The Impact of Imaging on Treatment
Endometrial cancer primarily presents at stage I (80% of cases), and the standard treatment is total abdominal hysterectomy and bilateral salpingo-oophorectomy. The clinical challenge is to effectively select patients at risk of relapse for more radical surgery (i.e., radical lymph node resection) and adjuvant treatment and to avoid overtreating low-risk patients. Furthermore, the recent introduction of laparoscopic techniques offers an alternative approach for patients who present with early disease [16, 17].
The major diagnostic factor necessary for the preoperative evaluation of endometrial cancer is to differentiate between stages IA and IB; this is becoming critical with increased use of hormonal treatment for stage IA disease in patients at high-risk for perioperative morbidity.
The risk of lymph node metastasis must be determined to select the appropriate surgical management. Differentiation of stage IB from stage IC has prognostic and morbidity implications. Stage IB patients should undergo lymph node sampling, whereas stage IC patients should undergo radical lymph node resection.
Gross cervical invasion requires preoperative radiation therapy or a different surgical plan—that is, radical hysterectomy instead of total abdominal hysterectomy.
Surgical staging of endometrial carcinoma is intended as the initial treatment and, at the same time, is used to identify patients who may require adjuvant therapy. The depth of myometrial invasion is probably the single most important morphologic prognostic factor because it correlates with tumor grade, tumor extension into the cervix, and the prevalence of lymph node metastases [18, 19]. The incidence of lymph node metastases (pelvic, paraaortic, or both) increases from 3% with superficial myometrial invasion (stage IB) to 46% with deep myometrial invasion (stage IC) [19–21]. Evaluation of the extent of myometrial invasion by gross inspection at surgery or at frozen-section analysis remains inaccurate in a significant proportion of patients [22, 23].
Controversy regarding the role of lymphadenectomy still exists, and practices vary considerably [24, 25]. Indications for lymphadenectomy include grade 1 or 2 tumors with deep myometrial invasion; all grade 3 tumors; cervical stroma invasion; and high-risk histologic subtypes, such as serous papillary and clear cell . Lymphadenectomy carries a high risk of complications and careful selection of high-risk patients is crucial for specialist surgical referral to a gynecologic oncology team .
In summary, MRI may assist in the preoperative assessment and surgical planning by accurately predicting the depth of myometrial invasion, cervical stroma invasion, lymph node involvement, and metastatic spread. MRI can also provide additional useful information such as uterine size, tumor volume, and ascites, and can reveal adnexal abnormalities that in turn may determine the surgical approach (i.e., transabdominal vs transvaginal vs laparoscopic). In high-risk patients due to comorbidity, MRI is useful in planning nonsurgical treatment options such as radiation therapy or hormonal therapy (stage IA).
Endometrial carcinomas are typically diagnosed at endometrial biopsy or dilatation and curettage, with MRI being reserved to evaluate the extent of disease. MRI is the most accurate imaging technique for the preoperative assessment of endometrial cancer because of its superb soft-tissue contrast resolution. The routine use of dynamic IV contrast enhancement is necessary for state-of-the-art MR evaluation of endometrial carcinoma [11, 13, 14, 26–29].
On unenhanced T1-weighted images, endometrial carcinoma is isointense to the normal endometrium. Although endometrial cancer may show high signal intensity on T2-weighted sequences, it is more typically heterogeneous and may even be of low signal intensity. After IV contrast medium administration, the normal inner myometrium shows avid enhancement earlier than the outer myometrium [13, 27, 28]. The maximum contrast between the inner and outer layers of the myometrium occurs at 50 seconds . In general, endometrial cancer enhances earlier than normal endometrium but later than the adjacent myometrium, allowing identification of small tumors, even those contained by the endometrium. The maximum tumor–myometrium contrast occurs during the equilibrium phase .
Imaging criteria for staging of endometrial cancer are based on the TNM or International Federation of Gynecology and Obstetrics (FIGO) classification system (Table 2). However, the FIGO staging system is based on surgical and pathologic findings alone, and imaging, although useful in the preoperative assessment of tumor stage, is not recognized as a method to be used for definitive staging.
TABLE 2: MRI Protocol for Imaging of Endometrial and Cervical Cancer
T1-weighted (upper abdomen and pelvis)
Evaluate bone marrow and lymph nodes
Evaluate lymph nodes and provide overview of pelvis
Visualize tumor and its relationship with myometrium
Axial oblique (short axis)
Evaluate depth of myometrial invasion in a second plane
T1-weighted (unenhanced and contrast-enhanced) 3D gradient-echo with fat saturation
Sagittal at 0, 1, 3, and 5 min Axial (short axis) at 4 min
Optimize the assessment of superficial (0 and 1 min), deep (3 min) myometrial invasion, cervical extension (5 min)
Optimize the assessment of depth of myometrial invasion (4 min) in a second imaging plane
T1-weighted (upper abdomen and pelvis)
Evaluate bone marrow and lymph nodes
T2-weighted with fat saturation
Evaluate lymph nodes and provide overview of pelvis (including hydronephrosis)
Visualize tumor and its extension to lower uterine segment, vagina, bladder, and rectum
Axial oblique (short axis)
Optimize the assessment of parametrial invasion in a second imaging plane
Note—T1-weighted can be conventional spin-echo, but in the interest of time, most centers use gradient-echo sequences, which are faster but at the expense of image quality. FSE = fast spin-echo.
MRI is significantly superior to sonography and CT in the evaluation of both tumor extensions into the cervix and myometrial invasion [30, 31]. The overall staging accuracy of MRI has been reported to be between 85% and 93% [13, 14, 26–30, 32, 33]. The routine use of dynamic IV contrast enhancement significantly improves the accuracy of the assessment of depth of myometrial invasion (accuracy of 55–77% for T2-weighted images vs 85–91% for dynamic contrast-enhanced images) [13, 14, 27, 29, 34].
Stage I—Stage I endometrial cancers include tumors confined to the uterine corpus. Stage IA tumors (limited to the endometrium) appear as normal or widened (focal or diffuse) endometrium. An intact junctional zone and a band of early subendometrial enhancement exclude deep myometrial invasion [13, 27] (Fig. 1A, 1B). Regardless of the MR sequence, the tumor–myometrium interface appears smooth and sharp. In stage IB disease (Fig. 2A, 2B), tumor extends less than 50% into the myometrium with associated disruption or irregularity of the junctional zone and a band of early subendometrial enhancement. If these landmarks are not present, stage IB tumor is suggested by an irregular tumor–myometrium interface. The presence of low-signal-intensity tumor during the equilibrium phase within the outer myometrium indicates deep myometrial invasion—that is, stage IC disease.
The sensitivity and specificity of MRI in the assessment of the depth of myometrial invasion range from 69% to 94% and from 64% to 100%, respectively [13, 14, 27, 29, 34]. An erroneous MRI assessment of the depth of myometrial invasion may occur when assessing a large polypoid endometrial carcinoma that distends the uterus so that the thin rim of myometrium is stretched over the carcinoma rather than showing deep infiltration [13, 21, 35]. Other causes include coexistent benign abnormalities (e.g., leiomyomas, adenomyosis) [15, 32, 35], congenital anomalies, indistinct zonal anatomy , poor tumor–myometrium contrast [15, 21, 29, 36–38], and tumor extension to the uterine cornu (Table 3).
TABLE 3: Classification of Endometrial Carcinoma Using TNM and International Federation of Gynecology and Obstetrics (FIGO) Staging Systems
Invasion less than or equal to half of the myometrium
Invasion of more than half of the myometrium
Invasion of cervix
Invasion of endocervical glands
Invasion of cervical stroma
T3 and/or N1
Local regional or local and regional spread
Involvement of serosa, adnexa, or both serosa and adnexa with or without positive peritoneal cytology
Metastatic to pelvic, paraaortic, or both pelvic and paraaortic nodes
Tumor extends outside pelvis or invades bladder or rectal mucosa
Invasion of bladder, bowel mucosa, or both
FIGO staging system is based on surgical and pathologic findings alone; imaging, althoug useful in preoperative assessment of tumor stage, is not recognized as a method for definitive staging.
Stage II—Stage II includes tumor extension beyond the uterine corpus into the cervix. In stage IIA, invasion of the endocervix appears as widening of the internal os and endocervical canal with preservation of the normal low-signal-intensity fibrocervical stroma (Fig. 3A, 3B).
Disruption of the fibrocervical stroma by high-signal-intensity tumor on T2-weighted images indicates cervical stroma invasion—stage IIB disease. Focal disruption of normal enhancement of the cervical mucosa by low-signal-intensity tumor on late dynamic contrast-enhanced MRI is useful in distinguishing cervical stroma invasion from polypoid tumor protruding from the endometrial cavity into the endocervix.
The accuracy of MRI in detecting cervical invasion reaches 92%, with sensitivities of 75–80% and specificities of 94–96% [14, 39].
Stage III—In stage III disease, tumor extends outside the uterus but not outside the true pelvis. Parametrial involvement—stage IIIA—appears as disruption of the serosa with direct extension into the surrounding parametrial fat. In stage IIIB disease, tumor extends into the upper vagina, and there is segmental loss of the low-signal-intensity vaginal wall. In stage IIIC disease, lymphadenopathy is present.
Stage IV—Stage IV disease is tumor that extends beyond the true pelvis or invades the bladder or rectum. The loss of low signal intensity of the bladder or rectal wall indicates stage IVA disease . In stage IVB disease, there is distant metastasis, malignant ascites, or peritoneal deposits. Peritoneal deposits are better seen on delayed dynamic contrast-enhanced MRI .
Sarcomas of the uterus are often highly malignant. They are rare, with an incidence of approximately 2 per 100,000 women over the age of 20 years, and account for 3–5% of all uterine cancers. The tumors are frequently large at the time of the examination, and it is difficult to determine the primary origin of the mass. MRI can provide an accurate preoperative assessment of uterine size and degree of involvement. The MRI features are nonspecific and may be indistinguishable from those of endometrial carcinoma [8, 41, 42]. However, uterine sarcomas tend to be large and heterogeneous with areas of hemorrhage and cystic necrosis. Deep myometrial invasion and peritoneal seeding are usually seen at presentation (Fig. 4A, 4B, 4C).
Leiomyosarcomas account for only 1.3% of uterine malignancies. Most leiomyosarcomas arise de novo from the myometrium, although malignant transformation of leiomyomas can occur. It has been suggested that an irregular margin of uterine leiomyoma may indicate malignant transformation [8, 43], but MRI cannot reliably differentiate between a leiomyoma undergoing benign degeneration and a leiomyosarcoma.
MRI of Malignant Neoplasms of the Uterine Cervix
The most common histologic type of cervical carcinoma is squamous cell carcinoma (90%) followed by adenocarcinoma (5–10%). Other rare histologic types include small-cell carcinoma, adenosquamous carcinoma, and lymphoma. MRI features of these rare tumors are the same as those of squamous cell carcinoma . However, small-cell carcinoma usually shows highly aggressive features, such as parametrial involvement, pelvic lymphadenopathy, and distant metastasis .
Carcinoma of the Cervix
Cervical carcinoma is the third most common gynecologic malignancy . In 2007, 11,150 new cases and 3,670 deaths are expected in the United States . Five-year survival rates vary between 92% for stage I disease and 17% for stage IV disease . During the past 50 years, there has been a steep decline in the number of deaths from cervical cancer. This improvement in mortality has been attributed to the development of the Papanicolaou test; only minor improvement has been achieved in the survival of invasive cervical cancer.
Patient preparation and coil choice for cervical cancer evaluation are similar to that for endometrial carcinoma. Although a body coil has been shown to provide similar staging accuracy, the use of a phased-array coil increases resolution and decreases imaging time [45, 46]. A basic imaging protocol should include axial T1-weighted spin-echo images with a large field of view and T2-weighted FSE images in the axial and sagittal planes with a small field of view (Table 1).
Cervical tumors are best seen on T2-weighted images. The sagittal plane allows evaluation of tumor extension into the body of the uterus and vagina. The axial oblique T2-weighted FSE sequence perpendicular to the long axis of the cervix is important in assessing parametrial invasion . Axial T2-weighted FSE imaging with fat saturation can be helpful in the evaluation of parametrial invasion, especially in younger women who have a prominent pericervical or vaginal plexus. Axial TI-weighted images of the abdomen are also included to identify enlarged abdominal lymph nodes.
The use of contrast medium is not necessary for cervical cancer examinations because it does not improve staging accuracy compared with unenhanced T2-weighted images [48, 49]. However, dynamic contrast-enhanced MRI may help distinguish recurrent tumor from postsurgical changes .
Impact of Imaging on Treatment
Staging of cervical cancer is still based on clinical FIGO criteria that—compared with surgical staging—can be erroneous in up to 32% of patients with stage IB disease and up to 65% of patients with stage III disease [50, 51]. The greatest difficulties in the clinical evaluation of patients with cervical cancer are the assessment of parametrial and pelvic sidewall invasion; accurate estimation of tumor size, especially if the tumor is primarily endocervical in location; and evaluation of lymph node metastases [52, 53]. Accurate pretreatment evaluation of these prognostic factors is crucial in determining appropriate therapy in patients with cervical cancer.
The most important issue in staging of cervical cancer is to distinguish early disease (stages I and IIA) that can be treated with surgery from advanced disease (stage IIB or greater) that must be treated with radiation alone or combined with chemotherapy. MRI is the best single imaging investigation that can accurately determine tumor location (exophytic or endocervical), tumor size, depth of stromal invasion, and extension into the lower uterine segment [44, 54–56]. MRI is accurate for evaluation of tumor size, usually within 0.5 cm of the surgical size, in 70–90% of cases [57–59]. Finally, MRI is useful in the evaluation of lymph node metastases .
MRI is recommended for evaluating patients with clinical stage IB disease or greater when the primary lesion is larger than 2 cm [44, 54, 55] because of a relatively high likelihood of parametrial invasion and lymph node metastases. Other MRI indications include evaluation of pregnant patients and patients with endocervical lesions .
On T1-weighted images, tumors are usually isointense to the normal cervix and may not be visible. On T2-weighted images, cervical cancer appears as a relatively hyperintense mass and is easily distinguishable from low-signal-intensity cervical stroma. On dynamic contrast-enhanced MRI, small tumors enhance homogeneously and earlier than the normal cervical stroma. Large tumors are frequently necrotic and may or may not enhance, but are often surrounded by an enhancing rim that facilitates tumor definition [60, 61].
The recommendations for diagnostic evaluation of tumor staging derive from the TNM and FIGO clinical staging systems (Table 4). In single-institution studies, MRI has been shown to be better than either CT or physical examination in depicting parametrial invasion [7, 62, 63]. The staging accuracy of MRI ranges from 75% to 96% [57–59, 62–65]. A recent prospective multi-center study conducted jointly by the American College of Radiology Imaging Network (ACRIN) and the Gynecologic Oncology Group (GOG) compared MRI, CT, and FIGO clinical staging in the pretreatment assessment of early invasive cervical cancer . The study showed that MRI was equivalent to CT for overall preoperative staging. However, MRI performed significantly better than CT for preoperative tumor visualization and determination of parametrial invasion. Reviewer agreement was higher for MRI reviewers than for CT reviewers .
TABLE 4: Classification of Cervical Carcinoma Using TNM and International Federation of Gynecology and Obstetrics (FIGO) Staging Systems
Invasive carcinoma identified only microscopically
Stromal invasion no greater than 3 mm in depth and no wider than 7 mm
Stromal invasion greater than 3 mm but less than 5 mm in depth and no wider than 7 mm
Clinical lesions no larger than 4 cm
Clinical lesions larger than 4 cm
Extension beyond cervix and involvement of the upper vagina (but not the lower vagina)
No parametrial invasion
Invasion of the lower third of the vagina, invasion extending to the pelvic sidewall, or both
Invasion of the lower third of the vagina
Extension to the pelvic wall or hydronephrosis (or both)
Invasion of the mucosa of bladder, rectum, or bladder and rectum; invasion extending beyond the true pelvis; or both
Spread to distant organs
FIGO staging system is based on surgical and pathologic findings alone; imaging, although useful in preoperative assessment of tumor stage, is not recognized as a method for definitive staging.
Stage I—Stage I tumors are confined to the uterus. Stage IA is defined as a microinvasive tumor that cannot be reliably shown on T2-weighted images. However, microinvasive disease may be detected on dynamic MRI as a strongly enhancing area on early arterial phase images . The accuracy in differentiating deep (> 3 mm) from superficial invasion has been reported to be 76%, 98%, and 63% on T2-weighted images, dynamic contrast-enhanced images, and contrast-enhanced T1-weighted images, respectively . Stage IB carcinoma appears as a high-signal-intensity mass in contrast to the low-signal-intensity fibrocervical stroma seen on T2-weighted images.
Young women with stage IA or small stage IB tumors who wish to retain their fertility are considered for trachelectomy, an operation that excises the cervix but preserves the uterine body and maintains fertility. MRI is accurate for predicting myometrial invasion by tumor and in showing the relationship of cervical carcinoma to the internal os with a sensitivity of 100% and specificity of 96% .
Stage II—In stage IIA tumors, segmental disruption of the upper two thirds of the vaginal wall without parametrial invasion is shown on T2-weighted images. Cervical stroma invasion (Fig. 5A, 5B) and tumor extension into the parametria are defined as stage IIB disease. The reported sensitivity of MRI in the evaluation of parametrial invasion is 69%, and the specificity is 93% [57–59, 62–65]. An intact low-signal-intensity cervical stroma virtually excludes parametrial invasion with a negative predictive value of 94–100% . Segmental disruption of the hypointense cervical stroma usually indicates full-thickness stromal invasion. However, additional features, such as a spiculated tumor–parametrium interface, soft-tissue extension into the parametria (Fig. 6A, 6B, 6C), or encasement of the periuterine vessels, are required to make a confident diagnosis of established parametrial invasion .
An important pitfall of MRI staging is overestimation of parametrial invasion on T2-weighted images in large tumors (accuracy of 70%) compared with small ones (accuracy of 96%) due to stromal edema caused by tumor compression or inflammation  (Table 3). This pitfall may lead to a higher rate of false-positive assessment of parametrial invasion in patients with large tumors, which must be considered when making the treatment decisions in these patients.
Stage III—In stage IIIA, vaginal involvement reaches the lower one third of the vaginal canal without extending to the pelvic sidewall. When the tumor extends to the pelvic sidewall (i.e., the pelvic musculature or iliac vessels) or causes hydronephrosis, it is defined as stage IIIB.
Stage IV—Once tumor invades the adjacent organs, such as the bladder and rectal mucosa, or distant metastasis occurs, the stage is defined as IV (Fig. 7A, 7B, 7C). MRI findings suggesting bladder invasion include focal or diffuse disruption of the normal low-signal-intensity posterior bladder wall, nodular or irregular bladder wall, mass protruding into the lumen of the bladder, or presence of bullous edema. Rectal invasion is rare and appears as segmental thickening and loss of the anterior rectal wall. Prominent strands between the tumor and the rectal wall may also indicate rectal invasion.
The reported sensitivity of MRI in the evaluation of bladder or rectal invasion is 71–100%, and the specificity is 88–91% [59, 64, 68]. The absence of bladder or rectal invasion can be diagnosed with sufficient confidence using MRI (negative predictive value = 100%) to safely obviate invasive cystoscopic or endoscopic staging in most patients with cervical cancer. This could potentially lead to a reduction in staging costs and morbidity . Although pelvic node metastases do not change the FIGO stage, paraaortic or inguinal node metastases are classified as stage IVB.
The Role of MRI in the Evaluation of Lymph Nodes in Uterine Malignancies
In patients with endometrial and cervical cancer, the presence of lymph node metastases suggests a poor prognosis, with a marked decrease in survival rates . For example, in surgically treated stages IB and IIA cervical cancer, survival rates decline from 85–90% to 50–55%, respectively, in the presence of metastatic lymph nodes . Lymph node involvement, which is not included in the FIGO staging system for carcinoma of the cervix, is also an important factor in the choice of adjuvant radiation therapy in both endometrial and cervical cancer. Surgical lymph node assessment is the gold standard for the diagnosis of lymph node metastases; however, lymphadenectomy carries a high risk of complications, and careful selection of high-risk patients is crucial for specialist surgical referral to a gynecologic oncology team . Therefore, from a clinical point of view, accurate preoperative assessment of lymph node metastases is very important in patients with endometrial and cervical cancer.
MRI and CT have comparable accuracies in detecting nodal metastases: 83–90% for CT and 86–90% for MRI [7, 62, 63, 70, 71]. They both rely on size criteria, which results in a low sensitivity (43–73% for MRI) due to the inability to identify metastasis in normalsize lymph nodes [5, 14, 64]. Recently, the use of lymph node–specific MRI contrast agents, such as ultrasmall superparamagnetic iron oxide (USPIO) particles, has been shown to improve the sensitivity and retain the high specificity of detection of lymph node metastases in patients with endometrial and cervical cancer . In their study, Rockall et al.  showed an increase in sensitivity from 29% using the standard size criterion (> 1 cm) to 93% using USPIO criteria on a node-by-node basis and from 27% to 100% on a patient-by-patient basis.
In cervical cancer, PET/CT has proved very valuable for lymph node staging with a sensitivity and specificity of 100% and 99.7%, respectively, for lymph nodes larger than 5 mm in diameter . Furthermore, 18F-FDG uptake within primary cervical cancer and lymph node metastases on FDG PET are reported as independent predictors of disease-free survival, which suggests that PET/CT may be the preferred imaging technique in patients with advanced carcinoma of the cervix for making treatment decisions, assessing nodal involvement, and determining prognosis .
The vagina is the sole site of recurrence in 30–50% of patients with endometrial carcinoma recurrence; the remaining patients develop pelvic or paraaortic lymph node involvement or systemic spread manifesting as hepatic, pulmonary, or osseous metastasis or peritoneal carcinomatosis. Manifestations of recurrent disease in cervical carcinoma can be characterized as typical and atypical. Typical manifestations involve the vaginal vault and lymph nodes. However, with the increasing use of pelvic irradiation in the treatment of this disease, less typical patterns of recurrence are becoming more frequent. These include peritoneal carcinomatosis and solid organ metastasis to the liver, adrenal gland, lung, or bone . Pelvic recurrence may involve other pelvic organs . Tumor extension into the bladder or rectal wall is suggested by abnormally high signal intensity on T2-weighted imaging.
Vaginal vault recurrence after radical surgery for endometrial or cervical cancer has similar appearances. This is indicated by loss of the low-signal-intensity linear configuration of the vaginal vault and visualization of an associated soft-tissue mass of high signal intensity on T2-weighted images, similar to that of the primary tumor (Fig. 8A, 8B).
In patients who have undergone radiation therapy, the critical issue is distinguishing recurrent disease from postirradiation changes. On MRI studies, recurrent disease appears as a heterogeneous mass on T2-weighted imaging, often similar to the appearance of the primary tumor  (Fig. 9A, 9B). However, T2-weighted images have low specificity for the detection of benign conditions, such as inflammation and edema, that cause increased T2-weighted signal. Dynamic contrast-enhanced MRI has been shown to be helpful in improving specificity and accuracy of tumor recurrence, with maximum tumor enhancement occurring between 45 and 90 seconds after contrast administration . However, early irradiation changes and the presence of infection continue to pose a problem because either may show enhancement. Serial imaging, imaging-guided biopsy, or PET may be required to further clarify the situation.
MRI, although not officially incorporated in the FIGO staging system, is already widely accepted as the most reliable imaging technique for the diagnosis, staging, treatment planning, and follow-up of both endometrial and cervical cancer. MRI protocols need to be optimized to obtain the best results and avoid pitfalls.
In endometrial cancer, MRI is reliable in predicting the depth of myometrial invasion and cervical extension, which correlate with the risk of lymph node metastases. Therefore, MRI is also valuable in selecting patients for lymph node sampling or lymphadenectomy who require specialist gynecology referral. MRI plays a central role in the evaluation of cervical cancer, primarily in identifying tumors without parametrial extension, thereby stratifying patients for surgery and radiation therapy. MRI also aids in the selection of patients for fertility-preserving surgery in early-stage disease and in the detection of recurrent disease after treatment. In summary, MRI plays a key role in staging, patient selection for treatment, and detection of disease recurrence.
Kinkel K, Ariche M, Tardivon AA, et al. Differentiation between recurrent tumor and benign conditions after treatment of gynecologic pelvic carcinoma: value of dynamic contrast-enhanced subtraction MR imaging. Radiology 1997; 204:55–63
Rockall AG, Sohaib SA, Harisinghani MG, et al. Diagnostic performance of nanoparticle-enhanced magnetic resonance imaging in the diagnosis of lymph node metastases in patients with endometrial and cervical cancer. J Clin Oncol 2005; 23:2813 –2821
Frei KA, Kinkel K, Bonel HM, Lu Y, Zaloudek C, Hricak H. Prediction of deep myometrial invasion in patients with endometrial cancer: clinical utility of contrast-enhanced MR imaging—a meta-analysis and Bayesian analysis. Radiology 2000; 216:444–449
Hawighorst H, Schoenberg SO, Knapstein PG, et al. Staging of invasive cervical carcinoma and of pelvic lymph nodes by high resolution MRI with a phased-array coil in comparison with pathological findings. J Comput Assist Tomogr 1998; 22:75–81
Lagasse LD, Creasman WT, Shingleton HM, Ford JH, Blessing JA. Results and complications of operative staging in cervical cancer: experience of the Gynecologic Oncology Group. Gynecol Oncol 1980; 9:90 –98
Bipat S, Glas AS, van der Velden J, Zwinderman AH, Bossuyt PM, Stoker J. Computed tomography and magnetic resonance imaging in staging of uterine cervical carcinoma: a systematic review. Gynecol Oncol 2003; 91:59 –66
Hricak H, Gatsonis C, Chi DS, et al. Role of imaging in pretreatment evaluation of early invasive cervical cancer: results of the intergroup study American College of Radiology Imaging Network 6651–Gynecologic Oncology Group 183. J Clin Oncol 2005; 23:9329 –9337
Peppercorn PD, Jeyarajah AR, Woolas R, et al. Role of MR imaging in the selection of patients with early cervical carcinoma for fertility-preserving surgery: initial experience. Radiology 1999; 212:395–399
Xue F, Lin LL, Dehdashti F, Miller TR, Siegel BA, Grigsby PW. F-18 fluorodeoxyglucose uptake in primary cervical cancer as an indicator of prognosis after radiation therapy. Gynecol Oncol 2006; 101:147 –151