Introduction

Cervical cancer is the third most common gynecologic malignancy in the United States. In 2011, 12,710 new cases and 4290 deaths¹ are expected in the United States. However, in the underdeveloped world where screening remains underutilized, cervical cancer is the second most common cancer in women, with 275,000 deaths worldwide in 2002.²

In the past 50 years, there has been a steep decline in mortality from cervical cancer that can be attributed to the use of the Papanicolaou (Pap) smear, one of the most effective and widely used screening tests. However, exfoliative cytology obtained with the Pap smear permits early detection of only early squamous cell carcinoma; this has led to a relative increase in the incidence of cervical adenocarcinoma.³

Epidemiology and Risk Factors

In the early half of the 20th century, cervical cancer was the leading cause of cancer-related mortality in women. In the developed world, this has declined significantly to its current position as the eighth leading cause of cancer-related deaths. This success has partly been due to the development of an effective screening test but also to the nature of cervical cancer in which the development of invasive cervical cancer is preceded by a long precancerous phase and the fact that these lesions, once detected, are then accessible and amenable to treatment.

There are numerous predisposing factors in the development of cervical cancer; however, early epidemiologic studies implicated an infectious agent as the most important factor. In 1983, this was identified as human papillomavirus (HPV). Although HPV infection is widespread, the majority of infections are cleared by cell-mediated immunity within 2 years and less than 10% of individuals will develop persistent infection. It is this persistent infection with HPV that plays a central role in the development of cervical cancer and can be identified in almost all cases. There are numerous HPV genotypes, of which HPV-16 and -18 have been determined as the most potent carcinogens. HPV-16 alone accounts for almost 60% of cervical cancers.³ In addition to its causative role in the development of squamous cell carcinoma of the cervix, HPV has also been implicated in the development of cervical adenocarcinoma and neuroendocrine carcinoma.

HPV targets the immature basal cells of the cervical epithelium; mature squamous cells, in contrast, are resistant to infection. In the cervix, there is a significant area of immature squamous metaplastic cells at the squamocolumnar junction. This is also called the transformation zone and anatomically is located where the squamous lining of the ectocervix meets the columnar cells of the endocervix at the level of the external os. This ring of metaplastic tissue is most susceptible to the carcinogenic effects of HPV infection.

In addition, numerous factors function in concert with exposure to high–oncogenic risk HPV infection and an inefficient immune response to increase the risk of developing cervical cancer, including

However, all these risk factors are linked to sexual behavior and infection with HPV and none, except for smoking, has been shown to be an independent risk factor. Smoking, even passive smoking, is associated with an increased risk of squamous carcinoma but not adenocarcinoma. The presence of cigarette carcinogens in the cervical mucus may be the explanation for this observation.

Anatomy and Pathology

Pelvic Anatomy

A basic overview of pelvic anatomy is important for staging cervical cancer because an understanding of pelvic ligaments, vessels, peritoneal reflections, and pelvic lymph node stations is vital in evaluating cross-sectional computed tomography (CT) and MRI.⁴

The important ligaments from the imaging perspective that are found in relation to the uterus, cervix, and ovaries include the broad ligament, round ligaments, uterosacral ligaments, cardinal ligaments, and suspensory ligament of the ovary (Figure 25-1).

Figure 25-1 Uterine anatomy.

The broad ligament is a peritoneal reflection that extends from the uterus to the pelvic sidewall. It contains the fallopian tubes along the upper margin; the cardinal ligament runs in the base of the broad ligament. It contains fat, connective tissue, the uterine and ovarian vessels, lymphatics, and the ovarian and round ligament. It is difficult to see except in the presence of ascites. The round ligament runs from the anterior wall of the uterus through the inguinal canal to the labia and is easy to see on CT. The cardinal ligament is an important structure that runs from the cervix and upper vagina to the obturator internus muscle. The uterine arteries run along the superior aspect of this ligament and help define this structure. As these arteries course from their origin from the internal iliac arteries to the edge of the uterus, they arch over the ureters creating the “arc sign,” which can frequently be seen on CT. The uterosacral ligament arises from the cervix and upper vagina to arc on either side of the rectum to the S2-S3 segments of the sacrum. This structure is thickened after radiation therapy. The suspensory ligament of the ovary carries the gonadal vessels to the ovary and is defined by the course of these vessels on cross-sectional imaging.

The main vascular supply to the uterus is the uterine arteries that arise from the internal iliac vessels, run along the superior aspect of the cardinal ligament, and then ascend on either side of the uterus to trifurcate and supply the fallopian tubes, fundus of the uterus, and ovaries (Figure 25-2). The ovarian arteries arise from the aorta below the renal arteries; the vaginal arteries arise from the internal iliac arteries.

Figure 25-2 Vascular supply to the uterus.

An important goal of imaging is to define the presence or absence of parametrial invasion. Consequently, the most important anatomic structures are those that assist in this evaluation. On CT and MRI, these structures are the ureters and the uterine vessels as they pass in the base of the broad ligament on either side of the supravaginal cervix. The ureters and uterine artery are in close proximity to either side of the supravaginal cervix, and abutment of the margins of ureters or uterine arteries is a confirmatory feature of parametrial invasion. As the extent of parametrial tumor progresses, ureters will get entrapped leading, to hydronephrosis. MRI has the added benefit of definition of the cervical stroma, which substantially increases the accuracy of this modality over CT for parametrial invasion.

Pelvic Nodal Anatomy

Disease spread to pelvic nodes is the most common pathway of tumor dissemination from the cervix. An understanding of the pathways of spread is essential for image analysis because these are the nodes that should be most closely scrutinized when evaluating cross-sectional imaging studies (Figure 25-3).

Figure 25-3 Pelvic nodes.

The nodal group most commonly first involved by tumor spread from pelvic tumors is the perivisceral nodes, which in the instance of cervical cancer are the parametrial nodes. This is followed by spread to pelvic sidewall nodes. Lymphatic spread from cervical tumors can spread to the pelvic nodes by three routes: (1) the lateral pathway of spread toward external iliac nodes, (2) the hypogastric route toward nodes along the internal iliac vessels, and (3) the posterior route, where lymphatics course along the uterosacral ligament to nodes along lateral sacral vessels and nodes anterior to the sacral promontory.⁵ The nodes along the external iliac vessels are subclassified into middle, medial, and lateral groups. The lateral chain nodes are located, as the name implies, lateral to the vessels; the middle chain nodes are between the external iliac artery and vein; and the medial chain nodes are located posterior and medial to the artery and vein. The nodes medial to the external iliac arteries are the group most commonly involved by metastatic spread from cervical cancer. These nodes are located in close proximity to nodes along the obturator vessels and are frequently grouped together with obturator nodes, although there is some controversy in this regard (Figure 25-4). All the pelvic nodal chains drain to the common iliac nodes. The common iliac nodes are also classified similar to external iliac nodes into middle, lateral, and medial subgroups. The middle subgroup is located posterior to the common iliac vein in the lumbosacral fossa, which is bordered posteriorly by the sacral vertebral body.⁶ This node is in close proximity to the L5 nerve root and can impinge on this root when enlarged, causing back pain (Figure 25-5). Spread from common iliac nodes is most commonly to the para-aortic nodes.

Figure 25-4 Bilateral obturator adenopathy (arrows) is identified on the axial T2-weighted fast spin echo magnetic resonance imaging (MRI) study. These nodes are seen to be fluoro-2-deoxy-D-glucose (FDG)-avid on the fused positron-emission tomography/computed tomography (PET/CT) images.

Figure 25-5 Common iliac adenopathy shows enlarged nodes involving the medial, middle, and lateral radial groups on the right (arrowheads). The right L5 nerve (arrow) root is in close proximity to the lateral chain nodes.

Key Points Anatomy, pathology, and epidemiology

• Persistent infection with HPV-16 and -18 is the most important risk factor in the development of cervical cancer.

• Cervical cancer has a long premalignant phase and, therefore, is accessible to early intervention and screening.

• Squamous carcinoma accounts for 85% of cervical tumors and arises from the zone located at the level of the external os and may present as exophytic or infiltrative masses.

• The uterine arteries and ureters as they course on the surface of the cardinal ligament define the level of the supravaginal cervix and site of parametrial invasion.

Clinical Presentation

In early cervical cancer, there may be no clinical symptoms as the disease progresses. Common clinical complaints are vaginal discharge, postcoital bleeding, irregular vaginal bleeding, or postmenopausal bleeding. In more advanced disease, lower abdominal or back pain or weight loss may also develop.

Patterns of Tumor Spread

The most common pathways of tumor spread in cervical cancer are direct invasion of contiguous pelvic structures and through lymphatics to lymph nodes; hematogenous spread is an uncommon pathway.

Direct spread occurs from penetration through the cervical stroma of the supracervical cervix into the parametrium, where the ureters and uterine arteries are located.

Lymph node spread in cervical cancer generally occurs in a sequential fashion with parametrial nodal involvement followed by external or internal iliac adenopathy, common iliac adenopathy, and then para-aortic adenopathy. It is very uncommon for para-aortic nodes to be involved in the absence of pelvic nodal disease.

As with all tumors, it is important to define regional and nonregional nodal groups; involvement of the latter upstages the tumor to stage IV because these nodes are viewed as M1 nodes. In the case of cervical cancer, parametrial, internal iliac, obturator, external iliac, common iliac, presacral, and lateral sacral are viewed as regional nodes, whereas inguinal, para-aortic, mediastinal, and supraclavicular are viewed as nonregional nodes, and consequently, qualify as metastatic disease.⁶

Staging

Unlike most other tumors, the staging of cervical cancer is primarily clinical, using the classification by the Federation of Gynecology and Obstetrics (FIGO) committee. This is because, in most countries where this disease is prevalent, elaborate cross-sectional staging techniques such as MRI, CT, and positron-emission tomography (PET)/CT are not available. Consequently, to maintain some uniformity between clinical trials across nations, the basis for initial staging is the clinical stage. The classification has undergone numerous revisions. This chapter references the most recent 2009 revision (Figure 25-6). Some important facts about the FIGO staging is that it applies only to squamous carcinoma; the clinical stage is determined prior to the start of therapy and cannot be changed because of subsequent findings once treatment is started. For evaluation of the T (tumor) stage, the following examinations are recommended: physical examination, preferably as an examination under anesthesia; colposcopy; cystoscopy; proctoscopy; intravenous urography; and chest x-ray. Suspected involvement of the rectal or bladder mucosa must be confirmed by biopsy. The tumor size plays an important prognostic role; consequently, subgroups have been created in both stages I and II for tumors less than 4 cm (T1a&b1 and T2a1) or greater than 4 cm (T1b2 and T2a2) in the tumor-node-metastasis (TNM) and FIGO classifications. It is noted that the definition of T categories corresponds to stages accepted by the FIGO classification (Table 25-1).7 A limitation of the clinical system of staging is that, compared with surgical staging, it can be erroneous in up to 32% of patients with stage IB and 65% of patients with stage III disease.^8,9

Figure 25-6 Stages I-IV of cervical cancer.

Table 25-1 TNM & FIGO Staging of Cervical Cancer

* FIGO no longer includes stage 0 (Tis).

† All macroscopically visible lesions—even with superficial invasion—are T1b/IB.

From Cervix uteri. In: Edge SB, Byrd DR, Compton CC, et al, eds. AJCC Cancer Staging Manual. 7th ed. New York: Springer; 2010:395–402.

Nodal status is not included in the FIGO classification because it cannot be assessed clinically; however, both the FIGO committee and the National Comprehensive Cancer Network (NCCN) guidelines encourage the use of additional modalities such as CT, MRI, and PET/CT to guide assessment of the nodal status and incorporation of the results into the treatment plan and prognostic data. The nodal status, however, is part of the TNM staging of cervical cancer.

The patterns of nodal spread have been described in a previous section; regional nodes for cervical cancer include obturator, internal iliac, external iliac, and common iliac.⁶ As previously mentioned, spread to para-aortic nodes constitutes metastatic spread. Another essential fact in assessment of the pathologic nodal status (pN) for the TNM classification is the number of nodes removed. At least six nodes must be available in the resected specimen; if the number of nodes is less than six, it qualifies as an NX, nodal status cannot be assessed.⁷

The definition of metastatic spread in the TNM or IVB in the FIGO classification includes invasion of bladder or rectal mucosa and distant spread. In the context of cervical cancer, distant spread most frequently presents as para-aortic, supraclavicular or mediastinal nodes, peritoneal spread, or involvement of lung, liver, or bones.

Imaging

As previously mentioned, the staging of cervical cancer is clinical; however, where cross-sectional imaging is available, it is used in treatment decision-making without changing the original clinical stage. The imaging modalities used include MRI, CT, PET/CT, and ultrasound.

A multi-institutional trial found that the overall accuracy of MRI and CT in the preoperative staging of early cervical cancer were comparable; however, in single institution studies, MRI is the most accurate technique for local staging of cervical cancer.¹⁰ The primary objective is to define tumors with parametrial invasion that precludes surgical resection and directs patients toward radiation and chemotherapy. MRI has proved to be the most effective technique in this regard with a high negative predictive value.

Primary Tumor

Magnetic Resonance Imaging

MRI is more useful than CT because of its superior soft tissue resolution. Its primary role is in determining tumor size, evaluating parametrial invasion, evaluating extent of uterine and vaginal involvement, and determining the involvement of adjacent pelvic structures. The patients who most benefit from an MRI are those with endocervical tumors, tumors larger than 2 cm at clinical examination, possible parametrial invasion, and pregnant patients.¹¹

The goal of MRI is to guide clinical decision-making. In this regard, parametrial invasion and pelvic nodal involvement are the most important aspects to be evaluated. The presence or absence of parametrial involvement determines whether surgical resection can be performed. Nodal involvement has significant prognostic value; however, also determines whether nodal dissection should be performed and the extent of radiation coverage if chemoradiation is the chosen treatment modality. In these different goals, MRI performs with variable accuracy. These are discussed because the limitations of a modality are important in its appropriate deployment.

Technique

A phased array coil is used to improve resolution and signal-to-noise ratio of the acquired images and is preferred over use of the body coil. Patient preparation can include fasting for 4 to 6 hours to decrease bowel peristalsis, which frequently degrades image quality on the sagittal T2-weighted images. Alternatively, frequency can be swapped to the anterior to posterior direction by adding no phase wrap (NPW) for small field of new images. In addition, vaginal gel may be used because it improves delineation of vaginal wall involvement, particularly the fornices.

The basic principles of high-quality MRI for staging of pelvic tumors are fundamentally similar whether staging cervical, endometrial, or rectal cancer (Table 25-2). These rely on small–field of view, high-resolution T2-weighted images obtained perpendicular to the plane of the tumor, preferably in two orthogonal planes. This entails using thin sections, preferably 3 to 4 mm, with a small field of view of 16 to 20 mm. Consequently, the most important sequence for evaluation of parametrial invasion is the oblique thin-section T2-weighted image obtained at right angles to the axis of the tumor, most importantly, the supravaginal portion of the cervix.

Table 25-2 Magnetic Resonance Protocol for Cervical Center

Protocol

Coronal T2-weighted images.

Axial T1-weighted images (large FOV).

Axial and sagittal T2-weighted images FOV 20-24.

Oblique axial T2-weighted images FOV 18-24 3 mm contiguous cuts.

Optional

Sagittal dynamic 3D LAVA images.

Postcontrast axial T1-weighted images.

It is recommended that the large FOV axial images be obtained from the level of the kidneys down through the pelvis to evaluate for retroperitoneal adenopathy and hydronephrosis.

FOV, field of view; LAVA, liver acquisition with volume acceleration; 3D, three-dimensional.

Intravenous contrast has limited value in the staging of cervical tumors. It has no role in evaluation of parametrial invasion; however, it can be used to assess involvement of adjacent structures such as the bladder. The dynamic images show small tumors as hypervascular lesions and have a reported high accuracy in determining depth of stromal invasion in small tumors.¹² However, larger tumors appear as centrally hypovascular with only marginal enhancement.

Evaluation of Images

Stage I

For tumors confined to the cervix, the smaller microinvasive tumors fall into category IA. In this subgroup, the role of imaging is limited, although there has been some work with dynamic contrast-enhanced MRI that identified tumors with greater than 4 mm depth of penetration into the stroma as hypervascular lesions.¹² The larger tumors that fall into the IIB category are seen on T2-weighted images as hyperintense masses that invade the hypointense stroma. However, in younger women, the stroma is occasionally hyperintense and the tumors may be harder to delineate. The size of tumors is an important prognostic indicator; this is reflected by the decline in 5-year survival rate from 84% to 66% in tumors larger than 3 cm.¹³ The larger tumors are also associated with an increased likelihood of nodal spread. The assessment of tumor size by MRI has an excellent correlation with size assessed by pathology.

Stage II

Stage II is defined by tumor spread outside the cervix into the upper third of the vagina or parametrium.

The most important information from the perspective of treatment planning is evaluation of parametrial invasion, the presence of which excludes surgical resection. In the assessment of parametrial invasion, MRI has a high negative predictive value (94-100%) with an intact dark stromal ring virtually excluding paramtrial invasion (Figure 25-7).^13,14 Disruption of the cervical ring generally indicates parametrial invasion, with additional signs such as spiculated tumor-parametrium interface and abutment of uterine vessels improving diagnostic confidence (Figures 25-8 and 25-9). An important limitation to keep in mind is that large tumors may be associated with peritumoral edema, which can be mistaken for parametrial invasion. This leads to a decline in staging accuracy from 90% for small lesions to 70% for larger tumors.^13,14 This important fact should be kept in mind at the time of image evaluation.

Figure 25-7 A 31-year-old with poorly differentiated adenocarcinoma. The oblique axial T2-weighted MRI study shows a hyperintense mass in the endocervical canal (small arrow) with an intact stromal ring (large arrow) excluding parametrial invasion.

Figure 25-8 A 44-year-old with poorly differentiated squamous carcinoma of the cervix presenting as an infiltrative mass involving the full thickness of the cervical stroma with bilateral parametrial invasion. Tumor encases both uterine vessels (thin blue arrows) and is occluding the left ureter (thick blue arrow). There are enlarged bilateral obturator nodes (yellow arrows).

Figure 25-9 A and B, A 53-year-old with cervical adenocarcinoma. Sagittal T2-weighted images show a hyperintense mass expanding the endocervical canal (arrow in A). The T2-weighted images orthogonal to the plane of the tumor show parametrial invasion on the left with a spiculated tumor-parametrium interface (blue arrow) and abutment of the left uterine artery (yellow arrow).

In terms of evaluation of the vaginal vault, disruption of the normal low signal of the vaginal wall by tumor is best identified on the sagittal T2-weighted images and reflects invasion. However occasionally with large tumors, the vaginal fornices may be stretched over the mass and it becomes difficult to assess for invasion. This can be overcome, to some extent, by the use of vaginal gel. However, in this circumstance, the information provided on MRI is not critical because it can also be obtained by visualization of the vaginal wall by clinical examination.

The assessment of invasion of the lower uterine segment can be obtained with MRI. Although it does not fall into information required for staging, it is important in young women who wish to retain fertility and are considering trachelectomy. MRI predicts the relationship of the tumor to the internal os with high accuracy.¹⁵

Stage III

In stage III, a vaginal involvement extends into the lower third of the vagina, and in IIIB, the extension is laterally until the pelvic sidewall. This lateral extension is generally associated with entrapment of the ureters and development of hydronephrosis. The MRI findings that suggest pelvic sidewall involvement include tumor within 3 mm of or abutment of the pelvic sidewall musculature.^16,17 MRI findings in vaginal involvement have been mentioned previously.

Stage IV

Invasion of adjacent pelvic structures such as the bladder or rectum or development of distant spread qualifies as stage IV disease. MRI has a high negative predictive value in excluding bladder or rectal involvement. However, the positive predictive value is lower. This is because abutment of the bladder or focal loss of the normal low signal of the bladder wall and development of hyperintense T2 foci along the anterior aspect of the posterior bladder wall do not always reflect tumor invasion. The sensitivity of bladder invasion is in the range of 71% to 100% and the specificity is in the range of 88% to 91%.^18,19 This limitation has led to an evaluation of dynamic MRI, and some studies have suggested that this leads to improved accuracy over evaluation of only T2-weighted images.²⁰ The more advanced cases of bladder invasion are relatively easier to diagnose as defined by nodular masses projecting into the bladder or development of a vesicovaginal fistula.

It is noted that recent multi-institutional studies have found both CT and MRI suboptimal in evaluating depth of stromal invasion, parametrial extension, and pelvic nodal metastasis. These studies also report lower accuracies for MRI staging than those seen in single-institution studies. However, the accuracy for assessment of tumor size and lower uterine invasion remains high, It is to be noted that the technique used in these studies did not incorporate high-resolution T2 scans or imaging in the orthogonal plane, which is important in the assessment of parametrial invasion.^10,21,22

Computed Tomography

CT has limited value in the assessment of local disease because 50% of the tumors are isodense to the cervical stroma on contrast-enhanced CT. CT is also suboptimal in its ability to distinguish tumor from normal parametrial structures leading to overestimation of parametrial involvement.²³

Key Points Imaging of primary tumor

• Parametrial invasion and nodal spread are the most important factors to be evaluated on imaging and determine whether treatment will be surgical versus nonsurgical and the extent of treatment.

• T2-weighted images obtained perpendicular to the cervical lumen are key to the evaluation of parametrial invasion.

Lymph Node Involvement

The presence and extent of nodal involvement are the most important prognostic factors in cervical cancer. In surgically treated cervical cancer, survival rates decline from 85% to 90% to 50% to 55% in the presence of nodes that are positive for tumor. In addition, the presence of metastatic adenopathy is important in treatment planning, defining radiation ports, and assessing need for and extent of surgical resection. It has been established in a multi-institutional study that the incidence of metastatic adenopathy in early cervical cancer stage IIA and below is in the range of 32%.²² The breakup of the likelihood of nodal metastasis on the basis of stage is less than 1% with stromal invasion less than 3 mm. However, in the presence of stromal invasion between 3 and 5 mm, this increases to 7%. In the larger tumors stage IB2, this increases to 20% to 25%.

Computed Tomography and Magnetic Resonance Imaging in Evaluation of Nodal Involvement

CT and MRI have comparable accuracies in the range of 83% to 85% in the assessment of nodal involvement. This is because both techniques rely on nodal enlargement as the criterion for malignant involvement. The limitation of both techniques is a low sensitivity in the range of 24% to 70%, which is due to the inability to detect metastasis in normal-sized nodes. However, although neither technique can distinguish inflammatory from malignant nodes, the specificity is reportedly high, in the range of 89% to 93%. It has been observed that there is an up to 27% incidence of necrotic adenopathy in cervical cancer.²⁴ The detection of necrotic nodes is more easily done with MRI. This finding, when seen, has a high specificity for the presence of metastasis. In an effort to improve accuracy in the assessment of nodal involvement, a variety of new MRI techniques are being evaluated, including the use of ultrasmall supraparamagnetic iron oxide (USPIO) particles and diffusion-weighted imaging. The use of USPIO particles has been reported to improve sensitivity on MRI from 29% to 93%; however, this agent is currently not available in the United States.²⁵ There is also some initial work to suggest that calculation of apparent diffusion coefficient (ADC) values on diffusion-weighted images improved sensitivity of nodal involvement; however, this needs to be more extensively evaluated.²⁶

At our institution, we believe that, in the assessment of nodal disease, it is important to incorporate the location of nodes into the evaluation of involvement. In other words, nodes along the anticipated pathways of spread such as the obturator, internal iliac, and common iliac should be scrutinized with a higher sensitivity than areas in which it is less common for spread to occur such as the lateral chain external iliac nodes. In addition, location of the tumor should also be incorporated into the assessment—for instance, if extension into the lower third of the vagina has occurred, the superficial inguinal nodes should be carefully assessed. Nodal morphology is also important in improving specificity and can be best assessed with the use of high-resolution (small–field of view/thin-section) images. The morphologic features most useful for assessing metastatic adenopathy are irregular or spiculated margins and, as mentioned previously, heterogeneity of the nodal signal characteristics.²⁷

Positron-Emission Tomography/Computed Tomography

In view of the low sensitivity of cross-sectional imaging in the evaluation of nodal involvement, efforts have been made to improve accuracy with the use of functional imaging modalities such as PET/CT.

The use of PET/CT has improved the accuracy of staging in that it has a high specificity in the detection of nodal metastasis. In addition, it improves sensitivity from the low sensitivities reported with MRI and CT to 83% to 91%; however, the problem with micrometastasis remains an issue. Metastatic nodes smaller than 5 mm are generally negative on PET.²⁸

The role of PET/CT in advanced cervical tumors is supported in a number of studies.^29,30 It establishes the presence of para-aortic adenopathy with high accuracy (Figure 25-10). It adds to the information available on CT, either by identifying nodes that were not enlarged on CT or by defining unexpected sites of disease. However, the value of PET in primary lymph staging is related to a high pretest probability of distant spread and is primarily in patients with locally advanced disease. Its role in early resectable cancer stages IA to IIA is questionable because the sensitivity has been found to be low, in the range of 10% to 53%.^31,32

Figure 25-10 A 31-year-old with poorly differentiated adenocarcinoma. A, Axial fused PET/CT scan shows FDG uptake in enlarged bilateral obturator nodes (blue arrows). There is significant FDG uptake seen in the primary cervical tumor (yellow arrow). B, A coronal maximum intensity projection (MIP) image shows extensive FDG-avid adenopathy in the retroperitoneum, along the common iliac and external iliac vessels.

Key Points Lymph node staging

• CT and MRI have comparable accuracies in the evaluation of nodal involvement because both rely on nodal size as the criterion for assessment.

• Newer techniques such as the use of USPIO and diffusion-weighted images have the potential to improve accuracy; however, this remains to be proved with more studies.

• PET/CT improves the assessment of nodal involvement in patients with advanced disease but has no value in early-stage cervical cancer.

Treatment

The treatment of cervical cancer is surgical in early stages (stage IA and B and IIA) and chemoradiation in more advanced stages.

The treatment of advanced cervical cancer is multidisciplinary. Women with malignant adenopathies, bulky stage IB2 tumors (primary tumor > 4 cm), and higher stages are at higher risk for local and/or distant recurrences. However, the significant decline in mortality of cervical cancer is related primarily to early diagnosis. In advanced disease, there has been no significant improvement in survival statistics.

New Therapies

Bevacizumab, a monoclonal antivascular endothelial growth factor antibody, and pazopanib, a vascular endothelial growth factor receptor inhibitor, have been evaluated in clinical trials and have increased the progression-free interval in patients with recurrent disease.^44,45 Epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors are ineffective in patients with refractory cervical cancer.⁴⁶ Cetuximab, a monoclonal inhibitor of the extracellular domain of EGFR, may have a modest efficacy in patients with refractory recurrence.^47,48

Surveillance

The majority of patients who develop recurrent disease do so within 2 years of treatment of the primary malignancy. Consequently, for this period, there should be close follow-up every 3 months. Subsequently, a 6-month evaluation can be obtained for the next 3 years, and after 5 years, an annual follow-up is standard of care.

In cervical cancer, half of the recurrences are within the pelvis; the most common sites are the vaginal cuff, cervix, parametrium, and pelvic sidewalls. Patients at high risk for recurrence or those who develop symptoms such as pain or vaginal bleeding should immediately be evaluated with imaging, particularly in the initial 2-year period. This can be with MRI, CT, or PET/CT. The goal remains the early detection of local disease and accurate characterization of local and distant spread to identify patients who may be eligible for a pelvic exenteration.

Traditionally, CT is the most widely used modality in the follow-up of patients. However, it is limited in distinguishing posttreatment change from tumor.⁴⁹ The presence of baseline posttreatment scans can somewhat obviate this problem by providing a comparison for follow-up CT scans. Any interval change identified on subsequent scans raises the possibility of recurrent disease.

MRI with superior soft tissue resolution has some advantages over CT, particularly when combined with dynamic scanning. The combination of T2-weighted images and early enhancement on dynamic scanning (<90 sec) has a high sensitivity (91%) but the specificity, although improved over evaluation with only T2-weighted images, remains low (67%).⁵⁰

Recent reports suggest that PET/CT detects recurrence earlier than CT or MRI.⁵¹ Studies using PET for surveillance of patients with no evidence of disease after primary treatment found a high sensitivity (90.3%) but a lower specificity (76.1%) for detection of recurrent disease. The recurrences were also detected earlier at 9 to 12 months than the historical average of 2 years, suggesting that this technique is more sensitive than previously used modalities and clinical examination. The additional value of PET/CT is the documentation of distant sites of disease, the presence of which could influence treatment decisions. The disadvantage of PET remains its low specificity and poor anatomic detail. Consequently, a reasonable strategy in the follow-up of patients with a high probability for recurrence would be to leverage the sensitivity of PET or MRI, supplemented by biopsy. If PET is used and identifies recurrence, an MRI of the pelvis should then be added in patients who are potential surgical candidates for pelvic exenteration or other radical surgeries. Alternatively, if MRI is the primary follow-up modality and if equivocal, a PET may be obtained to clarify the issue.