Improvements in the treatment of cervical carcinoma have made it possible to offer optimal and personalised treatment. Cervical cancer staging is based on clinical examination and histological findings. Many diagnostic methods are used in clinical practice. Magnetic resonance imaging is considered the optimal method for staging cervical carcinoma because of its high accuracy in assessing local extension of disease and distant metastases. Ultrasound has gained increased attention in recent years; it is faster, cheaper, and more widely available than other imaging techniques, and is highly accurate in detecting tumour presence and evaluating local extension of disease. Magnetic resonance imaging and ultrasound are often used together with computed tomography or positron emission tomography combined with computed tomography to assess the whole body, a more accurate detection of pathological lymph nodes and metabolic information of the disease.
Cervical cancer
Epidemiology
Cervical cancer is the second most common malignancy in women worldwide . Effective screening and prevention programmes in developed countries have resulted in a 75% decrease in incidence and mortality of cervical cancer over the past 50 years . In less developed countries, however, where screening and prevention programmes are not available, cervical cancer continues to be one of the most common causes of cancer-related morbidity and mortality among women .
Aetiology and risk factors
Invasive cervical cancer is related to age, with a mean age of 47 years at diagnosis in the USA . The major cause of cervical cancer is infection with high-risk human papillomavirus. The role of human papillomavirus in the cause of cervical cancer is tightly correlated with over-expression of two oncogenes (E6 and E7); their continuous expression is a necessary condition for the transformation of neoplastic cells . Other risk factors are represented by early onset of sexual activity, multiple sexual partners, and cigarette smoking.
Treatment
Treatment of invasive cervical cancer depends on clinical stage. Progress in therapeutic strategies allows tailoring of treatment: women with advanced disease will require neoadjuvant treatment (e.g. radiotherapy or chemotherapy alone or combined) possibly followed by surgery. In early disease, minimally invasive surgery is possible. The radicality of surgery needs to be tailored depending on tumour extension to minimise postoperative morbidity but also to preserve childbearing potential in young women.
Staging
The International Federation of Gynecology and Obstetrics (FIGO) recommends a clinical staging system for cervical cancer. It is well known, however, that the accuracy of such a system is suboptimal compared with surgical and pathological data. Tumour diameters, parametrial involvement, vaginal spread, infiltration of bladder wall, rectum mucosa, or both, hydroureter, hydronephrosis, and distant metastases are the fundamental parameters in the FIGO classification. Pelvic extension of cervical cancer can be assessed clinically by palpation, but other diagnostic examinations are required to assess distant metastasis or ureteral involvement.
In 20–30% of cases of early stage disease (defined on the basis of clinical examination), a significant discrepancy was reported between clinical staging and histological finding for tumour diameters and parametrial involvement . Cranio–caudal extension and parametrial infiltration are parameters that are not easy to estimate during the clinical examination; in particular, this is difficult in obese women. Indeed, FIGO staging does not include some of the tumoral parameters now recognised as significant prognostic factors, such as lymph-node metastases: 5-year survival of early stage cervical cancer (clinical FIGO stage) is 90% if no lymph-node metastases are present, but only 65% if they are . The degree of stromal infiltration, the distance between the upper limit of the cervical cancer, and the internal uterine orifice are other important prognostic factors .
Even if imaging techniques can provide information on significant prognostic factors, and is more accurate than clinical staging, clinical staging remains the only recognised gold standard worldwide. This is because clinical staging can also be used in low-income countries, where the prevalence of cervical cancer is greater than in high-income countries.
The National Comprehensive Cancer Network (NCCN) suggests that cross-sectional imaging techniques (computed tomography scan, magnetic resonance imaging [MRI], positron emission tomography [PET] combined with computed tomography), be used in stages equal to or higher than IB . Also, FIGO now encourages its use in the staging of cervical cancer for assessing prognostic factors, such as tumour size, parametrial and pelvic side wall invasion, adjacent organ invasion, and lymph-node metastases . Magnetic resonance imaging has been suggested as the optimal modality for staging cervical carcinoma FIGO stage IB1 or greater . Computed tomography has not proven accurate for assessing parametrial invasion or tumour size because of its limited contrast resolution . Whole-body PET-CT has now entered into clinical practice, in particular to define local extension and distant spread of cervical cancer, and to estimate its metabolic aspect.
Transvaginal ultrasound examination represents the gold standard examination for diagnosing gynaecological diseases; however, in the international guidelines, it is not included in the work up of cervical cancer, even if recent studies have shown a possible diagnostic role for this technique .
Imaging techniques for the evaluation of cervical cancer
Ultrasound
In the past 2 decades, ultrasound has gained increased attention in the assessment of cervical cancer . Ultrasound is faster, cheaper, and more widely available than other imaging techniques, and requires no preparation of the patient. Transrectal and transvaginal ultrasound give detailed images of the cervical tumour, because the probe is positioned close to the tumour, which means that high-frequency ultrasound can be used.
Fisherova et al. examined 95 women with transrectal ultrasound with histologically proven cervical cancer assigned to surgical treatment. The accuracy of tumour detection was 93.7% (95% CI 86.8 to 97.6), with a sensitivity of 93.4% (95% CI 85.3 to 97.8), and specificity of 94.7% (95% CI 73.9 to 99.8). In small tumours (<1 cm 3 ), the accuracy of tumour detection was 90.5% (95% CI 82.8 to 95.6), with a sensitivity of 72.0% (95% CI 50.6 to 87.9) and specificity of 97.1% (95% CI 90.1 to 99.6) . Similar accuracy for the detection of cervical cancer (93%, 95% CI 84 to 97) was obtained by Testa et al. who carried out transvaginal ultrasound examination in 68 women with histologically proven cervical cancer before surgery as primary treatment (33 women) or after neoadjuvant chemotherapy or chemoradiation . In a multicentre study including 182 women with early stage cervical cancer (FIGO IA2-IIA) scheduled for surgery (35% of them with cone biopsy before assessment), the transvaginal and transrectal ultrasound examination provided 96% accuracy for tumour detection, with 90% sensitivity and 97% specificity. The accuracy was not affected by cone biopsy before examination .
Ultrasound and morphological appearance of cervical cancer
At ultrasound examination, cervical cancer appears as a solid lesion with mostly hypoechoic echostructure compared with the surrounding cervical stroma ( Fig. 1 ). It rarely appears as isoechoic or hyperechoic lesion ( Figs. 2 and 3 ). In a multicentre study, involving 55 women with cervical cancer , isoechoic tumours were most commonly adenocarcinomas (13 out of 19 [68%]), whereas hypoechoic tumours were most often squamous cell carcinoma (11 out of 15 [73%]), ( P = 0.03).
At transvaginal and transrectal ultrasound examination, cervical cancers can be examined by pushing the probe against the uterus, whereby the tumour appears as an uncompressible lesion (Video 1) of different shape: according to its growth pattern, the cancer may appear as a ‘mushroom-shaped’ lesion when exophytic and ‘ovoidal’ or ‘conic-shaped’ lesions when endophytic ( Fig. 4 a,b and c). At colour or power Doppler examination, cervical tumours usually appear richly vascularised ( Fig. 5 ).
The following is the Supplementary data related to this article: Video 1
Longitudinal section of the uterus with cervical cancer. The forward movement of the probe facilitates the detection of a rigid, solid mass infiltrating the cervical stroma.
Ultrasound and assessment of the extension of the disease
After measuring three orthogonal diameters of the tumour ( Fig. 6 a and b), stromal infiltration must be assessed as involving less than or equal to two-thirds of the cervical stroma, more than two-thirds of the cervical stroma, or the whole cervical stroma. Extension up to the internal uterine orifice must be assessed. By positioning the endo-cavitary probe in the vaginal fornices, the vaginal tumour infiltration can be tested.
In the presence of cervical cancer infiltrating the full thickness of the cervical stroma, the extension into the anterior, posterior, and lateral parametria has to be assessed. The anterior compartment includes the bladder wall, vesico-vaginal septum, and anterior parametria (thin hyperecoic tissue located laterally to vesico-vaginal septum at longitudinal section through the pelvis). The posterior compartment includes the rectum, recto-vaginal septum, and posterior parametria (uterosacral ligaments). The lateral parametria correspond to the tissue located laterally to the cervix (paracervix).
Infiltration of the parametria is seen as irregular extension of hypoechogenic prominences into the pericervical tissue ( Figs. 7–9 ). Doppler ultrasound helps to distinguish between parametrial lateral tumour extension and hypoechogenic vessels . Infiltration of the vesicovaginal or rectovaginal septum is diagnosed if tumour extension is direct into the bladder or rectum, and the vaginal fornix is immobile against the bladder or rectal wall.
Data on the ultrasound assessment of disease extension are based on the results of only a few studies, and include women planned for surgery to treat clinically diagnosed early stage cervical cancer ( Table 1 ).
| Patients (N) | Examination approach | Ultrasound parameter | Sensitivity % (95% CI) | Specificity % (95% CI) | PPV (%) (95% CI) | NPV (%) (95% CI) | |
|---|---|---|---|---|---|---|---|
| Fisherova et al., 2008 | 95 | Transrectal | Tumour presence | 93.42 (85–98) | 94.74 (74– 100) | 98.61 (92–100) | 78.26 (56–92) |
| Tumour presence < 1 cm 3 | 72.00 (51–88) | 97.14 (90–100) | 90.00 (68–99) | 90.67 (82–96) | |||
| Parametrial infiltration | 83.33 (36–100) | 100.00 (96–100) | 100.00 (48–100) | 98.89 (94–100) | |||
| Testa et al., 2009 | 75 | Transvaginal | Tumour presence | 93 (84–97) | 50 (15–85) | 93 (90–96) | 50 (32–68) |
| Tumour invasion into >2/3 of the cervix stroma | 100 (81–100) | 75 (63–87) | 53 (44–62) | 100 (98–100) | |||
| Tumour vaginal infiltration | 33 (6–79) | 97 (93–100) | 33 (5–60) | 97 (95–99) | |||
| Parametrial infiltration | 60 (23–88) | 89 (81–97) | 30 (16–44) | 96 (93–99) | |||
| Vesico-vaginal septum infiltration | 100 (21–100) | 100 (77–100) | 100 (90–100) | 100 (99–100) | |||
| Recto-vaginal septum infiltration | – | – | – | – | |||
| Lymph node metastases | 9 (2–38) | 100 (73–100) | 100 (90–100) | 85 (81–89) | |||
| Epstein et al., 2013 | 182 | Transrectal or transvaginal | Tumour presence | 90 (85–95) | 97 (97–100) | ||
| Tumour presence (maximum diameter <2 cm) | 89 (83–95) | 89 (82–76) | |||||
| Tumour presence (maximum diameter > 4 cm) | 78 (64–82) | 99 (97–100) | |||||
| Tumour invasion into >2/3 of the cervix stroma | 88 (82–95) | 93 (88–98) | |||||
| Parametrial infiltration | 77 (54–100) | 98 (97–100) |
Transvaginal and transrectal ultrasound examination provides high sensitivity and specificity for the assessment of the depth of stromal invasion ; it provides high specificity and moderate sensitivity in evaluating parametrial involvement . The true diagnostic performance of ultrasound in detecting tumour growth in the vagina, infiltration of the vesico-vaginal septum, infiltration into the recto-vaginal septum, and lymph-nodes metastasis is impossible to estimate accurately, because of the lack of reported cases.
Ultrasound findings as prognostic factors
Colour Doppler results were found to be associated with prognostic factors in a study of 104 women with cervical carcinoma (stage IB-IIA) planned for primary surgery. Tumours with detectable blood flow had higher histologically proven vascular density, and were associated with higher risk of stromal and parametrial invasion and lymph nodal metastasis . In a study by Alcázar et al. , abundant blood flow (as assessed subjectively by the ultrasound examiner) was found more frequently in squamous carcinoma, moderately or poorly differentiated tumours, large tumours, and advanced stage tumours. Significantly lower resistance index and significantly higher peak systolic velocity were found in moderately or poorly differentiated tumours and advanced stage tumours. In 27 women undergoing primary surgery, abundant vascularisation at colour Doppler examination was associated with factors known to be predictive of recurrence (i.e. positive lymph nodes, parametrial and vaginal tumour infiltration, depth of stromal invasion, and lymph-vascular space involvement) .
Ultrasound and response to treatment
Vascular indices obtained at Doppler examination were examined as predictors of response to neoadjuvant treatment. Greco et al. in 1997 found that, in 10 women who successfully responded to chemotherapy, the Pulsatility Index and Resistance Index increased, and the maximum systolic flow velocity in tumour vessel decreased during treatment; in four non-responders, however, no change was observed in these parameters during treatment . In women with locally advanced cervical cancer, a lower number of vessels within the tumour, lower tumour vascular density (calculated as the ratio between the number of vessels and tumour volume), and higher Resistance Index in tumour vessel were found in women with complete response to preoperative chemoradiation than in those with partial response (response assessed in the surgical specimens) .
Magnetic resonance
Magnetic resonance imaging uses high-tissue contrast resolution, and is considered by many to be the optimal imaging method for evaluating the extension of cervical carcinoma because of its wide field of view and lack of ionising radiation. It is indicated to assess the extension of cervical cancer in clinical stage Ib1, and is particularly helpful in women who may benefit from fertility-sparing surgery. Moreover, MRI may be useful for monitoring the response to treatment, for detecting post-treatment complications, such as fluid collections, fistulae, and recurrence, but it can also be used as a template when external beam radiotherapy is planned.
The first pioneering studies on the use of MRI in preoperative staging of cervical carcinoma were carried out at the end of the 1980s and early 1990s . They showed extremely promising results in tumour recognition and in tumour volume estimation, despite rather disappointing accuracy in depicting extrauterine tumour extension and lymph-node involvement. The relevant technical improvements of MRI in subsequent years, such as the wide introduction of phased arrayed body coils and fast spin-echo sequences, enabled significant improvements to be made in its accuracy in evaluating extra-uterine extension of cervical carcinoma, confirming it superiority over computed tomography . From that time, MRI has been used routinely to assess the extent of cervical carcinoma . Over the years, pulse sequences and imaging protocols have been further developed and improved to reduce scanning times and to increase image quality. The inclusion of diffusion-weighted imaging (DWI) in abdomino-pelvic MRI protocols marked a new milestone in MRI of cervical carcinoma. Besides highlighting the presence of cervical carcinoma, DWI is promising as a marker for tumour aggressiveness and early treatment response .
Technical aspects
Magnetic resonance imaging of cervical carcinoma must be carried out on high field magnets (1.5 Tesla or more) using phase arrayed body coils. An antiperistaltic agent, such as butyl scopolamine bromide (Buscopan), should be intramuscularly administered, unless contraindicated (e.g. glaucoma), just before the examination to reduce bowel motion artifacts, and the bladder must be half full. If vaginal infiltration is suspected clinically, the vagina should be distended with sterile aqueous gel and the rectum should be distended with gel if rectal infiltration is suspected. The MRI protocol must always include non-fat-saturated small field of view and high-resolution fast spin echo T2-weighted images acquired according to the three planes orthogonal to the long axis of the cervical canal ( Fig. 10 ) (the so called ‘para-axial’, ‘para-coronal’ and ‘para-sagittal’ planes); these images are crucial for tumour detection and for evaluating local tumour spreads . To identify the presence of lymphadenopathy, fat-saturated fast spin echo T2-weighted images need to be acquired in the patient’s true axial plane from the femoral lesser trochanters to the renal hila. Non-fat-saturated T1-weighted images may be acquired in the patient’s true axial plane to recognise bone metastases, lymphadenopathy, and hematometra. Finally, diffusion-weighted images may be acquired using the same scanning parameters as above for high-resolution T2-weighted images to highlight the neoplastic tissue and to evaluate response after chemoradiotherapy. Intravenous contrast administration offers no significant advantages.
Magnetic resonance imaging appearance of cervical cancer
At MRI, cervical carcinoma may show different morphologic appearances, with endocervical, exophytic, or mixed growth. On T2-weighted images, cervical carcinoma almost always appears as an expansive or infiltrating inhomogeneous hyperintense (i.e. brighter signal intensity, ‘white’) mass compared with cervical stroma and adjacent structures ( Fig. 11 a). On diffusion-weighted images, cervical carcinoma appears as a markedly hyperintense mass with a corresponding hypointensity (i.e. darker signal intensity, ‘black’) on the apparent diffusion coefficient (ADC) map ( Fig. 11 b–c). On T1-weighted images, cervical carcinoma cannot be differentiated from adjacent structures. A cervical tumour that is clearly recognisable at MRI, and is evident on T2-weighted images, is usually at least a FIGO stage Ib .
Magnetic resonance imaging and detection of cervical cancer
Since the first published studies, MRI has always shown a high accuracy in the detection and measurement of cervical carcinoma, with a less than 5 mm discrepancy between the largest diameter measured on MRI, and the largest diameter measured at pathology testing in 93% of cases . Detection and evaluation of cervical carcinoma by MRI is typically carried out on the high-resolution T2-weighted images, which show the highest anatomical detail owing to their excellent intrinsic tissue contrast resolution. Tumour size must be evaluated on all the three orthogonal planes to recognise the largest diameter effectively. It is crucial to obtain precise measurements, because large tumour size can dictate treatment options; for example, fertility-sparing surgery is possible only for tumours less than 2 cm, whereas women with tumours greater than 4 cm should undergo neoadjuvant chemo-radiotherapy rather than radical surgery from the start. A potential pitfall in the evaluation of tumour size is the difficulty in discriminating between T2-hyperintensity, owing to peritumoural oedema, and T2-hyperintensity, owing to neoplastic infiltration, with a risk of over-staging; DWI may be helpful because of the absence of diffusion restriction in peritumoural oedema.
Magnetic resonance imaging and assessing disease extension
Cervical stromal infiltration
Cervical stroma appears as a hypointense ring on the T2-weighted images orientated in the para-axial plane. Neoplastic infiltration of cervical stroma can be recognised as the presence of T2-hyperintense expansive or infiltrating tissue within the hypointense ring ( Fig. 12 ). Superficial and deep infiltration can be differentiated by MRI. The finding of an incomplete interruption of the hypointense cervical stromal ring may be extremely helpful for excluding parametrial infiltration, with a negative predictive value of 94–100%. In cases of complete interruption of the hypointense cervical stromal ring, parametrial infiltration should be suspected even if a parametrial mass is not clearly visible .
Vaginal infiltration
The vagina appears as a virtual space surrounded by T2-hypointense walls on MRI. The vaginal fornices appear collapsed against the exocervix. Therefore, vaginal distension with sterile aqueous gel may help in distinguishing these structures. Vaginal infiltration is indicated by the presence of T2-hyperintense neoplastic tissue within the hypointense vaginal walls ( Fig. 13 ). The localisation of the infiltration should be described in detail (anterior or the posterior vaginal wall, confined to the upper two-thirds of the vagina (FIGO stage IIa) or also invading the inferior one-third of the vagina (FIGO stage IIIa). Magnetic resonance imaging has extremely high accuracy in evaluating the lower third of the vagina, but is less accurate than clinical examination in recognising minimal fornices involvement. In large exophytic tumours, vaginal infiltration may be overestimated .
Parametrial infiltration
At MRI, parametrial extension can be present only if the physiologically hypointense cervical stromal ring is completely interrupted by the neoplastic tissue. The first sign of parametrial involvement at MRI is the loss of the interface between the cervix and parametrial fat on para-axial images; indeed, in cases of early parametrial infiltration, the neoplastic tissue grows at the cervical-parametrial interface that becomes irregular, with a spiculated appearance ( Fig. 14 a). More advanced parametrial infiltration appears as a discrete soft tissue mass within the fatty parametrial tissue ( Fig. 14 b). Minimal parametrial infiltration must be accurately assessed because it often goes undetected at clinical examination. In the assessment of parametrial infiltration, MRI shows high accuracy (80–97%) and high negative predictive value (94–95%), with a specificity of about 94% and a sensitivity of 69%, if a women undergoing MRI staging is included .
Isthmic extension
The evaluation of isthmic extension, the measurement of the minimal distance between tumour and uterine internal orifice, and the measurement of cervical length, may be easily carried out at MRI ( Fig. 15 ) . In a recent systematic review, MRI has shown a specificity of 91% and a sensitivity of 97% in the evaluation of isthmic extension of cervical cancer, with an extremely high negative predictive value (99%) .
Pelvic wall involvement
Pelvic wall involvement can be accurately assessed at MRI, because MRI offers a panoramic view of the pelvis. Signs of pelvic wall infiltration are neoplastic tissue located at less than 3 mm from the pelvic wall, iliac vessel encasement, ureteral infiltration with upstream hydronephrosis ( Fig. 16 a and b), and infiltration of muscles with loss of their physiological hypointensity .
Involvement of adjacent structures
The presence of vesical, rectal wall infiltration, or both, can be accurately depicted by MRI, with a negative predictive value of up to 100%. More invasive investigations, such as cystoscopy and proctoscopy, can therefore be excluded . Vesical, rectal wall infiltration, or both, should be suspected if the fatty cleavage plane between the cervix and these structures appears completely obliterated. It is unequivocally diagnosed if neoplastic tissue grows within these structures ( Fig. 16 a and b, and Fig. 17 a and b). A signal intensity alteration of the physiologically hypointense rectal, vescical walls, or both, without gross neoplastic tissue growth, should be suspicious for infiltration, but may also be a consequence of bullous oedema; in these cases cystoscopy is mandatory .
Lymph-node involvement
The universally accepted MRI criteria for diagnosing metastatic pelvic lymph nodes is a shortest axis of the lymph node greater than 1 cm. As a consequence, micrometastases are easily missed. To increase the sensitivity of MRI in recognising lymph-node metastases, morphologic aspects may be added to the size of the nodes: rounded shape, inhomogeneous signal intensity, and spiculated margins are associated with malignant lymph nodes. The high reported accuracy values of MRI in recognising metastatic lymph nodes (about 90%) must be critically evaluated considering the low pre-test probability of nodal metastases; indeed, sensitivity values in the same studies vary between 40 and 60%, whereas specificity approaches 95% .
Treatment response
In cervical carcinoma re-staging after neoadjuvant chemoradiotherapy and brachytherapy, MRI plays an essential role in association with clinical examination. After chemoradiotherapy, cervical carcinoma usually shows a reduction in volume associated with a decrease in its T2-hyperintensity. Other common findings after chemoradiotherapy include urinary bladder and rectal thickening, usually associated with a diffuse hyperintensity on T2-weighted images, uterosacral ligaments thickening, presacral space widening, and hyperintensity of pelvic muscles on T2-weighted images. The main aim of MRI in this setting is the quantification of tumour shrinkage; poor tumour regression is associated with a higher risk of local recurrence . Partial response is characterised by a reduction in the largest diameter of the tumour of at least 30% compared with before treatment. Tumour progression is defined as an increase in the largest diameter of the tumour of at least 20%. Another extremely relevant parameter that needs to be evaluated after chemoradiotherapy is the extent of parametrial infiltration, unilateral compared with bilateral and medial compared with lateral ; indeed, it is important to compare laterality and extent of parametrial infiltration before and after chemoradiotherapy so that tumour response can be evaluated correctly.
Computed tomography
Computed tomography represents a widely available imaging modality that offers acquisition of high spatial resolution images during short scanning times. Unfortunately, computed tomography soft-tissue contrast resolution, even after the latest technological developments and also after the administration of intravenous iodinated contrast media, remains unsatisfactory. At computed tomography, cervical cancer appears isodense (the same density, i.e. as ‘gray’ as) compared with adjacent normal structures. Even large neoplasms may appear only as a non-specific cervical enlargement. Therefore, computed tomography is not indicated for local staging of cervical cancer. Gross tumour extension within parametria may be recognised also at computed tomography because of the presence of protrusions of soft tissue density within the fatty parametrial tissue or because of the presence of hydronephrosis. The sensitivity of computed tomography for the detection of parametrial infiltration is about 55% . Computed tomography may also recognise pelvic wall infiltration in case of presence of neoplastic tissue within 3 mm from the pelvic wall or in the case of iliac vessel encasement. Relatively recent work on the use of computed tomography for local staging of cervical carcinoma reported an overall staging accuracy of 53% .
In the detection of lymph-node metastases, computed tomography shows the same limitations as MRI. The only universally recognised criterion for diagnosing lymph-node metastases is a shortest axis of the lymph node greater than 1 cm. Computed tomography has a slightly higher sensitivity (60%) than MRI for detecting lymph-node metastases, maybe because of its higher spatial resolution .
Computed tomography remains the most accurate tool for recognising distant metastases, in particular to the liver and to the lungs, in advanced cervical cancers .
Positron emission tomography and computed tomography
Positron emission tomography combined with computed tomography (PET-CT) with the glucose analog 2-[fluorine 18] fluoro-2-deoxy-D-glucose (FDG) detects increased glucose metabolism associated with neoplastic lesions. It is recognised as an effective imaging tool in people with various malignant tumours, including cervical cancer. The association with computed tomography allows anatomic and functional information to be obtained in a single imaging session based on the structural changes visible on computed tomography together with the increased FDG uptake by tumour lesions provided by PET . The precise localisation of focal FDG uptake can be easily determined on fused PET-CT images. In uterine cervical cancer, the use of FDG PET-CT has spread widely. It is mainly used for initial staging and evaluation of treatment response. Other indications include radiation therapy planning, early detection of recurrence, and estimation of prognosis.
In 1978, FDG was developed, and it has become the most commonly used radiopharmaceutical for PET studies in oncology; its use for patients with known or suspected malignancies was approved by Food and Drug Administration in March 2000 . First experiences with FDG PET in women with cervical cancer go back to the end of the 1990s, showing promising results for detecting untreated and recurrent cervical tumours, as well as lymph-node metastases; however, FDG urinary excretion was problematic for image interpretation in some cases . The introduction of integrated PET-CT further enhanced the role of PET in cervical cancer, so that the data obtained with PET alone were reassessed with hybrid machines . The major advantage of PET-CT is its ability to provide metabolic and anatomic information by fusing images. These fused images facilitate localising lesions and differentiating between physiologic and pathologic FDG uptake in the abdomen and pelvis . Recently, integrated whole-body PET-MRI scanners have been introduced into clinical practice . Clinical studies are needed to clarify if PET-MRI will further improve the results in staging cervical cancer.
Technical aspects
Standard procedures are used to acquire FDG PET-CT scans . Images are most often acquired about 60 min after FDG administration; additional imaging 3 h after FDG administration may be helpful for detecting para-aortic lymph-node metastases or recurrent disease . The effective dose radiation in adults is 2 × 10 −2 mSv/MBq . To reduce radioactivity within the urinary bladder, the patient is asked to void completely before PET-CT; PET data are acquired in a caudal to cranial direction so that the pelvis is imaged first. Intravenous administration of 10–20 mg furosemide reduces radioactivity concentration in the urinary tract. Bladder catheterisation may be useful to facilitate evaluation of the uterine cervix and the vaginal cuff .
Interpretation of PET-CT scans qualitative and quantitative. Quantitative evaluation of areas of increased uptake is usually carried out by calculating standardised uptake value (SUV).
Physiologic FDG patterns
The knowledge of physiological FDG uptake is essential for the correct interpretation of pathological processes. In premenopausal women, functional ovarian FDG uptake may be seen around the time of ovulation (e.g. late follicular and early luteal phase) ( Fig. 18 ); endometrial uptake may be seen around ovulation and during menses .
FDG PET-CT for initial staging of cervical cancer
Most primary cervix cancers, with the exception of small lesions (<1 cm), show intense FDG uptake. This is related to GLUT-1 overexpression ( Fig. 19 ). Because of poor spatial resolution, however, FDG PET-CT is of limited value in the evaluation of tumour stage. It provides no information on parametrial invasion or involvement of adjacent organs. Both primary squamous cell carcinoma and adenocarcinoma show FDG uptake . Kidd et al. found that the mean SUV max was significantly higher for poorly differentiated than for well-differentiated tumours, and higher for squamous cell tumours than for non-squamous cell tumours . The intensity of FDG uptake in the primary tumour may have prognostic significance, as it is predictive of lymph-node involvement and disease outcome .
The role of FDG PET-CT in evaluating lymph-node metastates and for treatment planning is important as it is more accurate than computed tomography and MRI in assessing extra-pelvic disease ( Fig. 20 ). Tsai et al. found that 28% of women with untreated cervical cancer and MRI-defined pelvic node metastases, had their treatment plan modified by PET findings . False–negative findings may occur in the case of small lymph-node metastases (≤5 mm) and microscopic disease; overall, the false–negative rate of PET or PET-CT for metastatic para-aortic lymph nodes is 12% . False–positive results are explained by uptake in inflammatory nodes or by misinterpretation of physiologic activity in the bowel or urinary tract . Women with PET-positive lymph nodes have been reported to have significantly worse disease-specific survival than those with PET-negative lymph nodes . In a group of 83 women with positive FDG PET-CT lymph nodes, the lymph node SUV max was predictive of treatment response, risk of pelvic disease recurrence, disease specific survival, and overall survival .
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
