Ovarian cancer is the fifth most common malignancy in women and the most lethal among all gynecological diseases.1 Approximately 70% of malignant ovarian tumors are detected only at an advanced stage, which means that at the time of initial diagnosis, abdominopelvic dissemination has already occurred.1–6 The imaging findings most indicative of malignancy in ovarian cancer are peritoneal fluid, lymphadenopathies, and peritoneal carcinomatosis (PC).7–9
Ovarian cancer patient management depends on the results of initial staging, traditionally performed by laparotomy with simultaneous therapeutical “debulking” in case of PC. However, certain sites of the abdominal cavity are difficult to fully explore surgically.1,5 Furthermore, sampling errors may occur, leading to false-negative results and/or understaging in up to 30% of cases.4 Today, most clinicians would prefer a noninvasive staging to decide on neoadjuvant chemotherapy, if indicated, which permits downstaging before a surgical intervention.
Technical advances in multidetector CT (MDCT), MRI, and 18F-FDG PET/CT have enabled noninvasive staging, but which of these 3 techniques achieves the best patient management?
Several trials have already evaluated these imaging modalities for the initial detection of PC associated with ovarian cancer; however, they were done either not in comparison with each other10 or using old technical equipment for image acquisition and processing.2,3,8,11–15 Thus, we decided to undertake a reevaluation using modern image equipment. Our principal goal was to find out if the diagnostic value of these 3 modalities, either taken separately or together, would enable clinicians to noninvasively decide on whether to use neoadjuvant chemotherapy. Using surgical exploration with histopathology as the reference standard, we prospectively compared MDCT, MRI, and 18F-FDG PET/CT in ovarian cancer patients, in whom PC was suspected.
PATIENTS AND METHODS
After approval by our ethical committee, we prospectively studied 17 women, consecutively addressed to our department for primary ovarian cancer staging. In all women, concomitant PC was suspected, and debulking surgery was planned. Before imaging, each woman gave written informed consent.
Inclusion criteria were suspected ovarian cancer based on physical examination including an increased level of serum cancer antigen 125 (>35 U/mL) and/or sonographic findings of an ovarian mass with/without ascites. Exclusion criteria were known allergic reaction to iodinated contrast medium or gadolinium, renal failure (creatinine clearance <40 mL/min), and known contraindications to MRI.
We had to exclude 2 patients. In one, the final diagnosis was hepatobiliary cancer, and in the second, previous diagnostic laparoscopy lead to false-positive SUV uptake on 18F-FDG PET/CT. Thus, in the end, our study group comprised 15 women, all without surgery before imaging (mean age, 65 years; range, 31–89 years).
Technical Imaging Parameters
Shortly before surgery, all 15 women underwent MDCT, MRI, and 18F-FDG PET/CT. In 12 women, the 3 techniques were performed on the same day; in 1 patient, there was a delay of 1 day between FDG PET/CT and MDCT/MRI; in 1 patient, there was a delay of 4 days, and in 1 patient, there was a delay of 2 weeks. Thus, the mean interval was 1 ± 4 days, and the range was 0 to 14 days.
Multidetector CT was performed on a 64-row machine (Light Speed, VDT, 64 Pro; General Electric Healthcare, Milwaukee, Wis). After administration of a rectal enema of 1 L of water, we acquired axial slices (120 kV, 300 mA, 0.8 s/rotation, pitch 1.375, 2 mm/2 mm) from the diaphragm to the symphysis during portal venous phase at 70 seconds after IV iodine contrast medium injection (Iohexol, 300 mgI/mL, 3 mL/s, volume in milliliters = body weight + 30 mL), followed by multiplanar reconstructions with 30% overlap.
MR data were acquired on a 3.0-T MR scanner (TRIO or VERIO; Siemens Healthcare, Erlangen, Germany) with a maximum gradient strength of 45 mT/m. To include the whole abdomen in our examination protocol, we combined two 6-channel phased array body coils anteriorly and two 3-channel spine clusters posteriorly. After fasting for 6 hours before MRI and being administrated a rectal enema of 1 L of water, all patients were injected intravenously with an antiperistaltic agent (20 mg scopolamin butylbromide [Buscopan; Boheringer Ingelheim, Basel, Switzerland] or if that was contraindicated, 1 mg of glucagon [GlucaGen; Novo Nordisk, Basgvaerd, Denmark]).
The MR acquisition protocol included the whole abdomen. All our sequences were performed using the generalized autocalibrating partially parallel acquisition technique with an acceleration factor of 2. The axial plane was covered with 2 acquisitions of each sequence, centered on the upper and lower abdomen, respectively.
We started with axial and coronal breath-hold T2-weighted HASTE (TR, 1200 millisecond; TE, 89 millisecond; echo-train length [ETL], 256; number of excitations [NEX],1; matrix size, 320 × 240; section thickness/gap, 3/0.3 mm), followed by axial and coronal 3-dimensional VIBE (volumetric interpolation breath-hold examination) MR sequences (TR, 4.6 millisecond; TE, 1.7 millisecond; ETL, 1; flip angle, 9 degrees; NEX, 1; matrix size, 448 × 336; section thickness/gap, 4/0.8 mm) before and at 70 seconds after an IV gadolinium DOTA injection acquired in portal venous phase (Dotarem, 0.2 mmol/kg of body weight) followed by a 40-mL flush of 0.9% sodium saline.
Before IV gadolinium injection, we performed axial free-breathing fat-suppressed diffusion-weighted single-shot echo-planar MR sequences (TR, 6500 millisecond; TE, 66 millisecond; ETL, 1; receiver bandwidth, 1698 Hz/pixel; NEX, 3; matrix size, 168 × 126; section thickness/gap, 6/1.8 mm) while applying diffusion gradients in 3 orthogonal directions (section, phase, and frequency encoding directions), with increasing b values (0, 300, 600 s/mm2). The voxel size of the diffusion-weighted imaging (DWI) single-shot echo-planar sequence was 2.3 × 2.3 × 6.0 mm, and the acquisition time was approximately 7 minutes.
All patients were fasting for 6 hours or more and had a glucose plasma level less than or equal to 6.1 mmol/L before the 18F-FDG IV injection. PET/CT (Discovery LS; GE Healthcare Milwaukee, Wis) included a whole-body acquisition (from skull base to mid thighs) performed 70 ± 6 minutes after IV injection of 5.5 MBq/kg of 18F-FDG. PET acquisition was preceded by a craniocaudal unenhanced acquisition of MDCT (16-row detector) used for attenuation correction and localization (140 kV, 80 mA, pitch 1.5, 0.5 s/rotation, 5-mm slice thickness). PET data were subsequently reconstructed using an ordered subset expectation maximization method with 8 subsets and 2 iterations. A late PET/CT acquisition was also performed 104 ± 6 minutes after FDG injection and just after a bolus IV injection of an antiperistaltic agent (20 mg scopolamin butylbromide or 1 mg glucagon).
Reference standards were surgical exploration and histopathology. To allow for the best comparison between the 3 imaging modalities and our reference standards and for the best description of disease extension, we used the internationally recognized peritoneal cancer index proposed by Sugarbaker16 with a subtle modification. For our image analysis, we reduced the possible implant sites from 12 to 9 (Fig. 1), still covering the whole peritoneal cavity including the pelvis, but without differentiating the implants attached on the peritoneal surface from the ones attached to the bowel serosa. Thus, we simplified the comparison between our imaging modalities, but still assessed the exact topography of tumor extension (Fig. 1).
In addition, we included 3 sites of possible lymph node involvement, that is, retroperitoneal, iliac, inguinal, and basal pleural carcinomatosis. On each modality, possible ascites was also evaluated including the SUVmax of the pleural and abdominal fluid, when detected on PET/CT images.
Sugarbaker’s16 approach was to score not the number of peritoneal nodules on the MDCT and MR images, but the size of the largest implant detected in each quadrant by means of a 4-point grading system (Likert scale: LS0, no implant; LS1, implant ≤0.5 cm; LS2, implant ≤5 cm; LS3, implant >5 cm or confluent implants). In the case of any lymph node involvement, their small diameter was measured.
When analyzing the PET/CT images, we also took account of the 9 anatomical regions including the 3 lymph node sites, but without scoring the size of implants or of lymph nodes, instead measuring the SUVmax per quadrant and per lymph node site.
Blinded to all clinical information and independently, 1 radiologist (with 14 years of experience in abdominal imaging) read the anonymous MRI and MDCT images, whereas 1 nuclear physician (with 10 years of experience in PET/CT) read the anonymous PET/CT images. To reduce recall bias, the radiologist read the CT images 2 months later than the MR images. Each item was graded on a 5-point confidence scale (definitely absent, probably absent, undetermined, probably present, and definitely present).
In a joint reading session, both readers then compared MDCT, MR, and PET/CT images, still evaluating each quadrant separately on a lesion-by-lesion basis.
The same scoring system described previously was used by the operating gynecologist, who filled in the evaluation form during surgery, indicating the location and size of implants.
For statistical analyses, we used the software Stata 11.1. Sensitivity and specificity were evaluated for each technique. Each item was considered positive when evaluated with one of the 3 upper confidence levels. The χ2 test according to Pearson and receiver operating characteristic (ROC) curves with our 5-point confidence scale for calculating the area under the curve (AUC) of each technique were performed. Using Spearman rank correlation, we compared lesion sizes measured on MDCT and MRI. Finally, the interobserver agreement between MDCT and MRI was evaluated according to the κ statistics (κ = 0–0.20, slight; κ = 0.21–0.40, fair; κ = 0.41–0.60; moderate, κ = 0.61–0.80; substantial; and κ = 0.81–1, perfect agreement).17
All statistical differences were considered significant for P < 0.05.
The interval between imaging and surgery was 8.1 ± 2.4 days (range, 1–29 days). In all 15 women, ovarian cancer was histopathologically proven, and PC was associated in 10 (67%) of them, either stage III (n = 4) or stage IV (n = 6). Histopathology revealed 5 serous cystadenocarcinomas, 3 endometrioid adenocarcinomas, 3 poorly differentiated adenocarcinomas, 1 clear cell adenocarcinoma, 2 serous borderline tumors, and 1 mucinous borderline tumor. In 6 women (40%), the ovarian cancer involved both ovaries.
Altogether, we evaluated 135 abdominopelvic sites for PC and compared them with our reference standards.
For 74 anatomical sites (55%), PC was found, among them 13 (17%) with the largest implants measuring less than or equal to 0.5 cm, 40 sites (54%) measuring less than or equal to 5 cm, and 21 sites (29%) measuring greater than 5 cm.
Nine patients had ascites, among them 8 with PC (Fig. 2) and 1 woman without PC (Fig. 3). Pleural carcinomatosis was found in 3 patients (20%), and in 2 women (8%), the sigmoid colon was infiltrated.
Table 1 demonstrates our diagnostic results for each technique, that is, for the detection of PC as separate finding as well as for PC including pleural carcinomatosis and infiltrated lymph nodes. Figures 4 and 5 show the correspondent ROC analyses without any significant difference in the AUC between the 3 techniques (P = 0.12 and P = 0.11, respectively). Although there were no statistically significant differences, MRI had the highest sensitivity and negative predictive value and PET/CT had the highest specificity and positive predictive value, as well as accuracy and ROC AUC.
There was substantial agreement (κ = 0.68) between the interpretation of MDCT and MRI examinations with readings agreeing on 79% of the lesions.
Multidetector CT was more sensitive than MRI and PET/CT for detecting ascites with an AUC of 0.92 (95% confidence interval [CI], 0.75–1.0), 0.83 (95% CI, 0.50–1.0), and 0.83 (95% CI, 0.53–1.0), respectively, in ROC analysis; however, the sensitivity difference was not significant (P = 0.59).
According to Spearman rank correlation, the lesion size measured on MRI and MDCT compared with the histopathological results did not show any significant differences between the 2 techniques (MDCT 0.917 [0.89–0.94] and MRI 0.98 [0.84–0.90], respectively).
Sensitivity and specificity for detecting infiltrated lymph nodes were 77% (46%–95%) and 98% (87%–100%) for MDCT, 100% (75%–100%) and 98% (87%–100%) for MRI, and 93% (64%–100%) and 95% (83%–100%) for PET/CT, corresponding to an AUC of 0.88 (0.75–1.0), 1.0 (0.99–1.0), and 0.96 (0.88–1.0), respectively, in ROC analysis, which statistically means a trend in favor of MRI (P = 0.071).
For the detection of basal pleural carcinomatosis, there was a trend (P = 0.067) for differences in AUC in favor of PET/CT (MDCT 0.92 [0.75–1.0], MRI 0.67 [0.34–0.99], PET/CT 1.0 [0.99–1.0]).
Notably, PET/CT showed increased uptake (SUV 4.2 ± 1.1 g/mL; range, 3.4–5.4 g/mL) in thoracic lymph nodes in 3 patients (20%), which made us suspect metastases, but no histologic confirmation could be obtained. These supradiaphragmatic disease extensions were not detected on CT and were not investigated by MRI.
There is no universally accepted reference standard for imaging of PC.18 Our study did not reveal significant differences between MDCT, MRI, and PET/CT, but all sensitivity values were greater than 90%, as were the specificities (except for that of MRI). Preoperative imaging is crucial to determine the exact tumor extension. If PC, noninvasively assessed by initial staging, is too extensive for complete debulking, the women should be treated by neoadjuvant chemotherapy first. After downstaging, patients’ operability and clinical prognosis will be improved, as neoadjuvant chemotherapy has already been proven for the management of other abdominal malignancies, such as esophageal19 or rectal cancer.20
To the best of our knowledge, there has only been 1 prospective comparison of the 3 modalities, MDCT, MRI, and PET, performed simultaneously on 7 ovarian cancer patients. However, it focused on the ovarian tumors instead of highlighting PC.13
Table 2 summarizes the results of previous studies investigating 1 or 2 of our evaluated techniques for PC in ovarian cancer or in other abdominal malignancies.2,3,10–12,18,21–23 Their diagnostic values are mostly inferior or similar to ours (Table 2). Patient-based analyses yield higher diagnostic performance than site-based analyses, no matter the imaging modality. This certainly results from the frequently small size of single peritoneal implants and the subtle contrast difference with the surrounding anatomical structures.24
Multidetector CT, known for excellent spatial resolution, rapidity, robustness, and reproducibility of image acquisition, is today considered the workhorse of oncologic imaging. Unlike MRI, MDCT is particularly robust, when a large amount of ascites is present, as seen in 9 (60%) of our patients. Nevertheless, because 13 of our quadrants (17%) showed implants measuring less than or equal to 0.5 cm only, an excellent contrast resolution was required, which is the unique advantage of MRI.
The sensitivity of MRI for PC has been reported to increase by adding DWI MR sequences compared with using conventional MR sequences alone,9,25 provided that DWI is interpreted with the other acquired MR sequences.6,24,26,27 In our study, MRI turned out to be the most sensitive technique for PC, although without a statistically significant difference compared with MDCT and PET/CT. Our MRI results agree with those reported by Fujii et al26 evaluating MRI with DWI for PC in various gynecological malignancies.
Unlike MRI and MDCT, the decisive advantage of PET/CT is the whole-body coverage. Known as the modality of choice for detecting recurrent ovarian cancer,28–30 PET/CT is not yet routinely performed for initial ovarian cancer staging. PET with integrated MDCT is superior to PET alone for detecting PC because of better spatial attribution of focal radiotracer uptake.18,25,28,29 However, the limited spatial resolution remains an important issue.10 It may explain why in our study PET/CT showed a slightly lower sensitivity for PC than MDCT and MRI, albeit without reaching statistical significance. These results agree with those reported by Soussan et al,23 who directly compared PET/CT with MRI for PC arising from gastrointestinal malignancies. Thus, PET/CT can miss miliary peritoneal implants, especially when image misregistration due to respiratory and bowel movements occurs, or in patients with little mesenteric fat, in whom the intestinal loops are clustered together.
In some patients, MRI, and possibly also MDCT, leads to overstaging, which can be inferred from the lower specificity we obtained for MRI (84%) and MDCT (92%) compared with PET/CT (96%). This misinterpretation arise in presence of large quantities of ascites, in which the peritoneal vascularization prominently appears, thus rendering very difficult the exclusion of small peritoneal implants located between these serpiginous and often dilated vessels (Fig. 3). Therefore, we believe that in these cases, PET/CT is advantageous, especially if diagnostic aspiration of the peritoneal fluid cannot be performed. With PET/CT, massive ascites helps detect small peritoneal implants,15 possibly because of the greater distance among the different bowel loops, allowing for easier distinction between the physiologic intestinal activities from peritoneal implants attached to the bowel serosa.
Due to its whole-body coverage, PET/CT detected metastatic disease, not seen on MDCT or MRI in 20% of our patients. This may be particularly important because the detection of supradiaphragmatic disease means stage IV; these women would not benefit from optimal cytoreductive surgery and have shorter survivals.31
Shortcomings of PET/CT remain limited availability and higher costs, whereas the additional radiation exposure may not be an issue for these severely ill women. However, the initial higher costs for PET/CT may be counterweighed by more straightforward patient management after PET/CT, especially in case of clinically unexpected supradiaphragmatic disease extension that would be discovered by PET/CT. Indeed, the cost effectiveness would be best studied in large patient populations, for instance using existing national oncology PET registries.32,33
Our study has limitations. First, the daily equipment schedule in our department required us to use a 3.0-T instead of a 1.5-T MR magnet, despite important quantities of ascites in some women. Because this large amount of intraperitoneal fluid is a highly conductive medium, 3.0-T magnets present more, or even new, artifacts compared with 1.5-T scanners. This hampers the diagnostic quality of MR images.34 However, MRI, and in particular DWI, has been proven feasible at 3.0 T in advanced ovarian cancer.35
Unlike other authors,25 we deliberately refrained from a sequential analysis of our MR images (first without DWI sequences, then including them). We think that DWI has now become an integrated part of abdominal oncological imaging protocols and should be analyzed in conjunction with the other MR sequences. We also refrained from measuring the apparent diffusion coefficient, mainly because of the small size of many peritoneal implants.
We also performed only non–contrast-enhanced PET/CT to compare MDCT versus PET/CT. It is possible that contrast-enhanced PET/CT might have better performance characteristics than MDCT and PET/CT, but this was not assessed here.
Finally, although our study population was quite small, we believe our results represent valid findings for a single center. We would like to confirm them by a larger multicenter trial.
In conclusion, our study yielded no significant differences between MDCT, MRI, and PET/CT for detecting PC in ovarian cancer patients. MRI was the most sensitive technique, and PET/CT was the most specific one. Thus, MDCT, known as the fastest, most economical, and widely available modality, may be the examination of choice, if only one can be performed. If MDCT is negative or inconclusive, PET/CT or MRI may offer additional insights. PET/CT, as a whole-body modality, may provide more accurate preoperative evaluation of supradiaphragmatic metastatic extension.
1. Bristow RE, Duska LR, Lambrou NC, et al. A model for predicting surgical outcome in patients with advanced ovarian carcinoma using computed tomography. Cancer
. 2000; 89: 1532–1540.
2. Ricke J, Sehouli J, Hach C, et al. Prospective evaluation of contrast-enhanced MRI
in the depiction of peritoneal spread in primary or recurrent ovarian cancer
. Eur Radiol
. 2003; 13: 943–949.
3. Tempany CM, Zou KH, Silverman SG, et al. Staging of advanced ovarian cancer
: comparison of imaging modalities—report from the Radiological Diagnostic Oncology Group. Radiology
. 2000; 215: 761–767.
4. Forstner R. Radiological staging of ovarian cancer
: imaging findings and contribution of CT and MRI
. Eur Radiol
. 2000; 17: 3223–3246.
5. Pannu HK, Bristow RE, Montz FJ, et al. Multidetector CT of peritoneal carcinomatosis
from ovarian cancer
. 2003; 23: 687–670.
6. Nougaret S, Addley HC, Colombo PE, et al. Ovarian carcinomatosis: how the radiologist can help plan the surgical approach. Radiographics
. 2012; 32: 1775–1800.
7. Sohaib SA, Sahdev A, Van Trappen P, et al. Characterization of adnexal mass lesions on MR imaging. AJR Am J Roentgenol
. 2003; 180: 1297–1304.
8. Pannu HK, Horton KM, Fishman EK. Thin section dual-phase multidetector-row computed tomography detection of peritoneal metastases in gynecologic cancers. J Comput Assist Tomogr
. 2003; 27: 233–340
9. Iafrate F, Ciolina M, Sammartino P, et al. Peritoneal carcinomatosis
: imaging with 64-MDCT
and 3 T MRI
with diffusion-weighted imaging. Abdom Imaging
. 2013; 37: 616–627.
10. De Iaco P, Musto A, Orazi L, et al. FDG-PET/CT in advanced ovarian cancer
staging: value and pitfalls in detecting lesions in different abdominal and pelvic quadrants compared with laparoscopy. Eur J Radiol
. 2011; 80: e98–e103.
11. Coakley FV, Choi PH, Gougoutas CA, et al. Peritoneal metastases: detection with spiral CT in patients with ovarian cancer
. 2002; 223: 495–499.
12. Forstner R, Hricak H, Occipinti KA, et al. Ovarian cancer
: staging with CT and MRI
. 1995; 197: 619–626.
13. Kubik-Huch RA, Wörffler W, von Schulthess GK, et al. Value of (18
F)-FDG positron emission tomography, computed tomography, and magnetic resonance imaging in diagnosing primary and recurrent ovarian carcinoma. Eur Radiol
. 2000; 10: 761–767.
14. Axtell AE, Lee MH, Bristow RE, et al. Multi-institutional reciprocal validation study of computed tomography predictors of suboptimal primary cytoreduction in patients with advanced ovarian cancer
. J Clin Oncol
. 2007; 25: 384–389.
15. Yoshida Y, Kurokawa T, Kawahara K, et al. Incremental benefits of FDG positron emission tomography over CT alone for the preoperative staging of ovarian cancer
. AJR Am J Roentgenol
. 2004; 182: 227–233.
16. Sugarbaker PH. Surgical responsibilities in the management of peritoneal carcinomatosis
. J Surg Oncol
. 2010; 101: 713–724.
17. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics
. 1977; 33: 159–174.
18. Dirisamer A, Schima W, Heinisch M, et al. Detection of histologically proven peritoneal carcinomatosis
with fused 18
. Eur J Radiol
. 2009; 69: 536–541.
19. Ajani JA, Komak R, Putman JB, et al. A three-step strategy of induction chemotherapy then chemoradiation followed by surgery in patients with potentially resectable carcinoma of the esophagus or gastroesophageal junction. Cancer
. 2001; 15: 279–286.
20. Reerink O, Verschueren RC, Szabo BG, et al. A favourable pathological stage after neoadjuvant radiochemotherapy in patients with initially irresectable rectal cancer correlates with a favourable prognosis. Eur J Cancer
. 2003; 39: 192–195.
21. Kim WH, Won KS, Zeon SK, et al. Peritoneal carcinomatosis
in patients with ovarian cancer
. Enhanced CT versus 18
F-FDG PET/CT. Clin Nucl Med
. 2013; 38: 93–97.
22. Klumpp BD, Schwenzer N, Aschoff P, et al. Preoperative assessment of peritoneal carcinomatosis
: intraindividual comparison of 18
F-FDG PET/CT and MRI
. Abdom Imaging
. 2013; 38: 64–71.
23. Soussan M, Des Guetz G, Barrau V, et al. Comparison of FDG-PET/CT and MR with diffusion-weighted imaging for assessing peritoneal carcinomatosis
from gastrointestinal malignancy. Eur Radiol
. 2012; 22: 1479–1487.
24. Kyriazi S, Collins DJ, Morgan VA, et al. Diffusion-weighted imaging of peritoneal disease for noninvasive staging of advanced ovarian cancer
. 2010; 30: 1269–1285.
25. Satoh Y, Ichikawa T, Motosugi U, et al. Diagnosis of peritoneal dissemination: comparison of 18
F-FDG PET/CT, diffusion-weighted MRI
, and contrast-enhanced MDCT
. AJR Am J Roentgenol
. 2001; 196: 447–453.
26. Fujii S, Matsusue E, Kanasaki Y, et al. Detection of peritoneal dissemination in gynecological malignancy: evaluation by diffusion-weighted MR imaging. Eur Radiol
. 2008; 18: 18–23.
27. Bozkurt M, Doganay S, Kantarci M, et al. Comparison of peritoneal tumor imaging using conventional MR imaging and diffusion-weighted MR imaging with different b
values. Eur J Radiol
. 2011; 80: 224–228.
28. Pannu HK, Bristow RE, Cohade C, et al. PET-CT in recurrent ovarian cancer
: initial observations. Radiographics
. 2004; 24: 209–223.
29. Fulham MJ, Carter J, Baldey A, et al. The impact of PET-CT in suspected recurrent ovarian cancer
: a prospective multi-centre study as part of the Australian PET data collection project. Gynecol Oncol
. 2009; 112: 462–468.
30. Markman M. The impact of PET-CT in suspected recurrent ovarian cancer
: a prospective multi-centre study as part of the Australian PET data collection project. Author reply. Gynecol Oncol
. 2009; 114: 536.
31. Hillner BE, Sieger BA, Shields AF, et al. The impact of positron emission tomography (PET) on expected management during cancer treatment. Cancer
. 2009; 115: 410–418.
32. Hillner BE, Siegel BA, Liu D, et al. Impact of positron emission tomography/computed tomography and positron emission tomography (PET) alone on expected management of patients with cancer: initial results from the national oncologic PET registry. J Clin Oncol
. 2008; 26: 2155–2161.
33. Risum S, Høgdall C, Loft A, et al. Does the use of diagnostic PET/CT cause stage migration in patients with primary advanced ovarian cancer
? Gynecol Oncol
. 2010; 116: 395–398.
34. Merkle EM, Dale BM. Abdominal MRI
at 3.0 T: the basics revisited. AJR Am J Roentgenol
. 2006; 186: 1524–1532.
35. Sala E, Priest AN, Kataoka M, et al. Apparent diffusion coefficient and vascular signal fraction measurements with magnetic resonance imaging: feasibility in metastatic ovarian cancer
at 3 Tesla. Eur Radiol
. 2010; 20: 491–496.