The most award winning
healthcare information source.
TRUSTED FOR FOUR DECADES.
Abstract & Commentary
Synopsis: Dynamic enhanced helical CT and MR imaging both have high accuracy but only moderate sensitivity in detecting pelvic lymph node metastases of cervical carcinoma using a size criterion of maximal axial nodal diameter larger than 10 mm. More optimal criteria for maximal axial nodal diameters may be 9 mm for helical CT and 12 mm for MR imaging.
Source: Yang WT, et al. Comparison of dynamic helical CT and dynamic MR imaging in the evaluation of pelvic lymph nodes in cervical carcinoma. Am J Radiology 2000;175:
In patients with cervical carcinoma, the presence of metastases in pelvic lymph nodes has important implications for prognosis and the choice of appropriate therapy. Unfortunately, the reported sensitivities and specificities of current radiologic techniques for distinguishing normal and metastatic pelvic lymph nodes have generally been less than optimal. In this study, Yang and associates in Hong Kong performed a prospective evaluation of two newer dynamic enhanced techniques, helical CT (with 7-mm section thickness) and MR imaging (with 4-, 5-, and 6-mm section thickness), for making this distinction in 43 women with biopsy-proven cervical carcinoma. They recorded the minimal and maximal axial diameters, CT attenuation, MR signal intensity, and areas of necrosis in visible pelvic lymph nodes. These findings were correlated with the results of pathologic examination of 949 lymph nodes surgically removed from 76 hemipelves in these women. Sixty-nine (7%) of the lymph nodes, from 17 hemipelves (22%), contained metastatic tumor at pathologic examination. Using a maximal axial diameter of more than 10 mm or the presence of central necrosis as criteria for nodal metastasis, Yang et al found the respective sensitivities of CT and MR imaging on a hemipelvis basis to be 64.7% and 70.6%; specificities, 96.6 % and 89.8%; positive predictive values, 84.6% and 66.7%; negative predictive values, 90.5% and 91.4%; and accuracies, 89.5% and 85.5%. Then, using receiver operating characteristic curve analysis, they determined that optimal size criteria for distinguishing benign from metastatic nodes in their patient group were maximal axial nodal diameters of 9 mm for CT and 12 mm for MR imaging. Central necrosis within a node had a positive predictive value of 100% for metastasis; as would be expected, necrotic nodes tended to be large (with a mean maximal nodal diameter of 2.3 cm at CT and 1.9 cm at MR imaging). Nodal shape, MR signal intensity, and CT enhancement pattern could not reliably be used to distinguish benign and metastatic nodes.
COMMENT BY DAVID M. PANICEK, MD
Radiologists are well aware of the major limitations of lymph node assessment at CT and MR imaging: normal-sized nodes may contain microscopic deposits of metastatic tumor, whereas nodes free of tumor may be enlarged in reaction to some other process. Various approaches have been undertaken in an attempt to improve upon this situation, including analysis of parameters such as nodal shape, nodal enhancement, and site-specific and disease-specific size criteria for nodal enlargement; and use of node-specific contrast materials. To date, the only clinically useful sign of pelvic lymph node metastasis is nodal enlargement. The choice of a specific cutoff size beyond which a node is considered enlarged is controversial, however, and results in an unavoidable tradeoff between sensitivity and specificity; the smaller the size criterion used, the higher the sensitivity for nodal metastasis—but at the expense of lowered specificity. Comparison of various studies of lymph nodes is complicated by the use of short-axis diameter in many, long-axis diameter in others, and lack of any specification of how the node was measured in a few; also, use of different slice thicknesses and types of cross-sectional imaging techniques further confuses the issue.
As in all studies, there are some limitations of this one. Only patients who underwent surgery were enrolled, so some patients with very advanced disease were excluded; therefore, the study population was biased toward patients with less advanced disease. The section thickness was different for CT and MR imaging, which may at least partly account for the different optimal cutoff size criterion found for each technique. Most importantly, no node-by-node correlation between imaging and pathology was performed; thus, an imaging test would be considered "correct" if any node in a hemipelvis was scored as positive for metastasis on the images—even though that node may not have corresponded to any pathologically proven metastasis-bearing node surgically removed from elsewhere in the hemipelvis.
This study represents a good step in the right direction for refining the radiologic criteria for pelvic lymphadenopathy in patients with cervical carcinoma, but additional research is needed before new size criteria (or other parameters) can be adopted for routine clinical use.