Enter Your Email Address to Receive a
Copy of MedicExhange Member Demograhpics

A semi-automated system developed at Washington University in St. Louis combines two types of computer algorithms previously only used separately to process data from computerized CT scans of the lungs.

Lung cancer presents a special challenge to clinicians attempting to evaluate the effectiveness of radiation treatment and determine the total dose of radiation received by the tumor and surrounding tissues. The reason is simple: lung tumors change position as an individual breathes during medical scans. This unavoidable movement of the lungs makes it difficult to accurately assess tumor volume (particularly in the very small malignant nodules that are more treatable if detected early) and track any changes in size that may have resulted from treatment.

Issam M. El Naqa, an assistant professor of radiation oncology at Washington University in St. Louis, and his colleagues have devised a novel solution. Their semi-automated system combines two types of computer algorithms previously only used separately to process data from computerized tomography (CT) scans of the lungs.

So-called deformable regression algorithms are used to create a consistent set of coordinates on which tumor position and size can be mapped over the course of treatment, and segmentation algorithms allow tumors to be precisely located and distinguished from other lung tissue (or 'segmented') in CT images. El Naqa and his colleagues, realizing that "both approaches could significantly benefit from the results of the other algorithm if coupled in the same framework," created a new program that does just that.

El Naqa, who has tested the combination algorithm in a preliminary study of four people with non-small cell lung cancer, says that the method provides more accurate and consistent results for tracking tumor changes. He says the technique "would allow us to learn more about tumor response to treatment and potentially be used in treatment adaptation," or, perhaps, in the pre-planning of treatment strategies that would reduce the overall levels of toxic radiation received by people undergoing radiotherapy for lung cancer.

In related work, researchers at the University of California, San Diego, have developed a computer algorithm to localize the position of lung tumors during fluoroscopic imaging. Fluoroscopic imaging permits clinicians to view images obtained in real time, but because of the poor contrast between lung tumors and normal lung tissue, tumors can be essentially invisible. However, the tumors may move in concert with anatomic features that are easier to visualize and that can serve as stand-ins.

"The algorithm that we are developing will be able to automatically select surrogate anatomic features whose motions are correlated with tumor motion," says study senior author Steve Jiang, Associate Professor and Director of Research in the Department of Radiation Oncology at UCSD. "Thus, by tracking their motion we can derive the positions of the unseen tumors."

The findings were presented at the 50th meeting of the American Association of Physicists in Medicine (AAPM), July 27-31, in Houston, Texas.