Ultrasonic Detection of Metastases in Dissected Lymph Nodes of Cancer Patients
Current histological methods can miss micrometastases (< 2.0 mm) in dissected lymph nodes because nodes are cut into sections that are at least 2-mm thick for examination, and the entire node volume cannot be evaluated microscopically. In this study, high-frequency, quantitative ultrasound (HFU, QUS) methods were applied to freshly dissected lymph nodes to detect micrometastases based on their microstructural properties. 3-D ultrasound data were acquired from 40 nodes from 22, colorectal-cancer patients using a single-element, 25 MHz transducer. Significant cancer was detected subsequently in 7 of the 40 nodes. Node images were segmented semi-automatically in 3-D, and echo signals were processed to yield basic spectral parameters (slope, intercept, and midband) values plus QUS estimates associated with tissue microstructural properties (scatterer size and acoustic concentration). Images were formed by expressing local QUS estimates as color-encoded pixels and overlaying the color on conventional, gray-scale ultrasound images. Linear discriminant analysis classified nodes based on intercept, midband, size, and acoustic concentration. ROC methods assessed classification performance. 3-D QUS images interactively displayed spectral-parameter and QUS values. Linear-discriminant methods produced an area under the ROC curve of 1.000 based on size and intercept; interestingly, the areas for size alone and for intercept alone were 0.986. These initial results appear to validate spectrum-analysis-based QUS methods for distinguishing cancerous from non-cancerous tissue in lymph nodes. The Areas under the ROC curves suggest that this approach can be valuable clinically to identify nodal micrometastases that current histologic methods can miss.
KeywordsColorectal cancer Quantitative ultrasound High-frequency ultrasound Spectrum analysis Lymph nodes Metastasis Micrometastasis Maximum likelihood
The authors acknowledge the support provided by NIH grant CA100183 and the Riverside Research Institute Biomedical Engineering Research Fund.
- 4.Feleppa, E.: Ultrasonic tissue-type imaging of the prostate: implications for biopsy and treatment guidance. Cancer Biomark. 4(4–5), 201–212 (2008)Google Scholar
- 5.Coron, A., Mamou, J., Hata, M., Machi, J., Yanagihara, E., Laugier, P., Feleppa, E.J.: Three-dimensional segmentation of high-frequency ultrasound echo signals from dissected lymph nodes, Proceedings of the 2008 Ultrasonics Symposium, edited by K.R. Waters (Institute of Electrical and Electronics Engineers, Piscataway, 2008), pp. 1370–1373CrossRefGoogle Scholar
- 9.Tateishi, T., Machi, J., Feleppa, E.J., Oishi, R., Jucha, J., Yanagihara, E., McCarthy, L.J., Noritomi, T., Shirouzu, K.: In vitro diagnosis of axillary lymph node metastases in breast cancer by spectrum analysis of radio frequency echo signals. Ultrasound Med. Biol. 24(8), 1151–1159 (1998)CrossRefGoogle Scholar
- 10.Mamou, J., Coron, A., Hata, M., Machi, J., Yanagihara, E., Laugier, P., Feleppa, E.: High-frequency quantitative ultrasound imaging of cancerous lymph nodes. Jpn. J. Appl. Phys. 48, 07GK08-1 (2009)Google Scholar