CT and MR are Equally Accurate and Adequate to Stage Advanced Ovarian Cancer
Abstract & Commentary
Synopsis: Advanced ovarian cancer can be staged with high accuracy by either computed tomography or magnetic resonance imaging. Ultrasound’s lesser accuracy was primarily due to poorer results in imaging peritoneal metastases in the subdiaphragmatic spaces. Whether this limitation can be overcome was not tested in this study and is not yet known.
Source: Tempany CM, et al. Staging of advanced ovarian cancer: Comparison of imaging modalities—Report from the Radiological Diagnostic Oncology Group. Radiology 2000; 215:761-767.
Ovarian cancer is the second most common gynecologic cancer and causes the most deaths. A "silent killer," the five-year survival rate for all stages is 46%, but the survival varies significantly by stage: 93% for localized disease, but 25% for advanced disease. Less than one-quarter of patients present with local disease. Many cases of higher stage disease are only discovered at laparotomy. Accurate preoperative staging by imaging would be of value to assess potential sites for surgical biopsy, to ensure a proper surgical approach to debulking, and to plan referral for appropriate chemotherapy.
Under the auspices of the National Cancer Institute of the National Institutes of Health, the Radiological Diagnostic Oncology Group (RDOG) performed a prospective, multi-institutional assessment of the accuracies of magnetic resonance (MR) imaging, computed tomography (CT), and gray-scale Doppler ultrasonography (US) in the diagnosis and staging of ovarian cancer. Overall diagnostic accuracy for malignancy was not significantly different among the three modalities.1 Now the group presents an analysis of subdata reflecting key staging sites in patients with advanced disease—Stages III and IV—precisely those patients in whom detailed therapy planning is needed. (In the 1986 staging system1 of the International Federation of Gynecology and Obstetrics [FIGO], patients with peritoneal implants outside the pelvis or positive retroperitoneal or inguinal lymph nodes are classified in Stage III while those with distant metastases including malignant pleural effusions or parenchymal liver metastases are Stage IV.)
Tempany and colleagues compared the accuracy of detection of spread of ovarian cancer to 11 individually identified sites in the peritoneum, 10 separate lymph node sites, and the liver parenchyma (lesions on the liver surface were included with the peritoneum). All patients underwent at least two of the three imaging examinations within four weeks of the surgical staging and, when appropriate, debulking. All examinations were performed on high-quality imaging equipment using standardized protocols including intravenous iodinated radiographic contrast material for CT studies and 1.5 Tesla MR systems with imaging before and after intravenous administration of gadolinium-based contrast material. Images were interpreted in a standardized fashion prospectively at each institution by a single (but different) radiologist for each modality. Radiologists rated their degree of suspicion of malignancy at each site. Receiver operating characteristic (ROC) curve statistical analysis was performed for the imaging findings using the results of surgery and histology as proof. The area under the ROC curve, Az, indicates overall diagnostic accuracy.
Of the 280 patients who met the study criteria, 118 had malignancy and 73 (62% of patients with malignancy) had Stage III or IV. Metastases to the peritoneum occurred in 70 patients, to lymph nodes in 20, and to the liver in seven.
For peritoneal metastases, MR and CT were more sensitive than US (95%, 92%, and 69%, respectively), but US was more specific than MR or CT (93%, 80%, and 82%, respectively). ROC curve analysis showed MR and CT superior to US (Az = 0.96, 0.96, and 0.86, respectively). (Some of these comparisons had P values of 0.05, technically not statistically significant at the P < 0.05 level typically used as the criterion.) Subgroup analysis showed that the subdiaphragmatic spaces were a key area where CT and MR were superior to US.
For lymph nodes, the Az for MR, US, and CT were 0.76, 0.68, and 0.57, respectively. The values for MR and CT were significantly different; the values for US did not differ significantly from those of either MR or CT. For the liver, the small number of metastases limited the analysis, and overall accuracies as indicated by Az did not differ significantly among the three modalities.
Comment by James H. Ellis, MD
In the spirit of full disclosure, I must state that I was one of the 15 radiologists who interpreted images for this study and, for a short while, a small amount of my salary was supported (but not supplemented) by my being a co-investigator on this project. I was a co-author of a different paper based on this patient population.1
The current study relies heavily on complex statistical analysis for its results. The work was complicated by the logistical inability to obtain all three imaging modalities (MR, CT, and US) in each patient; many patients underwent only two of the three modalities. The result is that pairwise comparisons were not always made on precisely the same populations of patients.
Regardless of these logistical problems, some useful conclusions can be drawn. In patients with advanced stage ovarian cancer due to peritoneal, lymph node, or hepatic metastases, peritoneal metastases are most common as the defining cause of the advanced stage. In the current study of 73 patients with advanced disease, 70 had peritoneal metastases compared to 20 with lymph node metastases and seven with liver metastases. Thus, any imaging technique that is to be used to stage advanced ovarian cancer needs to be able to visualize peritoneal metastases. MR and CT were superior to US in this regard, especially due to the ability of MR and CT to detect subdiaphragmatic peritoneal metastases. Whether now knowing that the subdiaphragmatic region is a weakness of US could lead to more effort and thus better results or whether this is a noncorrectable limitation of US technique is unknown.
Perhaps surprisingly, none of the techniques had high sensitivity in the detection of metastatic lymph nodes (41%, 39%, and 32% for CT, MR, and US, respectively; slightly different numbers are given in a table compared to in the text). Lymph node sizes were not reported; many metastases may have been within nonenlarged lymph nodes. Ultrasound was a relatively close third in sensitivity, compared to what one might typically expect of ultrasonographic evaluation of lymph nodes.
Another finding that might not have been predicted was the success of MR in identifying peritoneal metastases and abdominal (e.g., porta hepatis) lymph node metastases compared to CT even in the absence of bowel contrast agent for MR. Whether an MR oral contrast agent would further improve the accuracy of MR remains to be tested.
Some limitations of the study should be noted. The examinations were performed from 1993 to 1996. All three imaging modalities have undergone technical improvements, particularly CT with multidetector technology (some of the exams were performed in dynamic mode, others in helical mode). The CT technique was chosen to favor the pelvis over the abdomen: patients were scanned from the symphysis pubis upward. Especially for scanners in dynamic mode, this means that the liver was not imaged during the optimal phase of intravenous contract enhancement. The number of patients with liver metastases was too small (7 patients) to draw any conclusions, perhaps justifying this unorthodox technique. With current faster CT scanners, a technique that does not sacrifice abdominal image quality for pelvic image quality could be used.
MR and CT have high accuracy in staging advanced ovarian cancer and either modality is appropriate. Local factors, such as availability, cost, and patient preference or contraindications, may be important in choosing which one to use. Ultrasound is surprisingly accurate given the need to assess peritoneal implants and lymph nodes. The results of this study could lead to investigation of how ultrasound technique might be improved to overcome its circumscribed deficiency in detecting subdiaphragmatic peritoneal metastases.
Reference
1. Kurtz AB, et al. Diagnosis and staging of ovarian cancer: Comparative values of Doppler and conventional US, CT, and MR imaging correlated with surgery and histopathologic analysis—Report of the Radiology Diagnostic Oncology Group. Radiology 1999;212:19-27.
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content