|
|||||||||||||||
|
Home
|
More Data —
Computer-Aided Detection Tries to Quantify Wash-in/Wash-out Principle At the last RSNA meeting, 3TP Imaging Sciences rendered its own name obsolete. The company had advanced beyond its eponymous 3TP (three-time-point) algorithm and into the more complex full-time-point (fTP) algorithm for computer-aided detection (CAD) of breast and prostate cancers. As such, the White Plains, N.Y.-based organization changed its name to CAD Sciences to better reflect the direction it was taking for breast, prostate, and therapy monitoring on MRI and for other applications such as angiography and CT. Previous to this latest development, the 3TP analysis itself was used with patients receiving contrast-agent injection during an MRI examination to diagnose breast or prostate cancer. In a nutshell, the CAD algorithm compared signal intensities between three different time points to create a colorized map from the MR images that helped determine whether tissue was benign or malignant. Users were able to view cellular dynamics within static images, allowing them to discern subtle differences in tissue biology. The newer fTP algorithm provides more diagnostic information, as it encompasses as many as time points as available to create colorization that’s based on tumor biology as well as the signal changes. Because it uses additional time points, signal problems at any point have less impact on the overall result. Business Foundation The company’s 3TP postprocessing software produced accurate visual representations of the presence and patterns of contrast-induced enhancement on MR data sets, enabling radiologists to view cell dynamics in static images that, in turn, helped them discern subtle tissue differences in benign and malignant lesions. Essentially, the 3TP algorithm provided vital information that was previously difficult to obtain with conventional mammography and breast MRI by placing a color-coded map on top of each original MR image slice. The added color more clearly differentiates what the algorithm labels as benign or malignant tissue. This additional information helps physicians make more effective treatment decisions. However, the company, perceiving the inherent limitations of using just three time points, sought to improve on the method, a creative compulsion that fostered the development of fTP and necessitated the company’s name change. In the past year, CAD Sciences has moved forward with fTP, offering the only products on the market comprised of the algorithm. However, it hasn’t completely forsaken 3TP. “The 3TP method is available for customers who have become comfortable using it,” reports CAD Sciences Chief Technology Officer Henry W. Wyszomierski, “but the fTP method, in moving beyond three time points, enables the user to generate more robust information and enter into more advanced applications such as monitor response to therapy.” That information, he says, involves permeability and extracellular volume, key physiological markers for cancer that can lead to better lesion characterization and, ultimately, more effective patient treatment. The 3TP Method The three recorded time points, taken from a dynamic series of MRI scans, include a single precontrast and two postcontrast benchmarks processed by the algorithm. Time-point selection is determined by how quickly the contrast enters the breast lesion (wash-in) and how quickly it leaves (wash-out). “When talking about the 3TP method, some people use the term wash-in/wash-out curve analysis. It is the general method used in the industry,” says Wyszomierski. This wash-in/wash-out profile methodology drives the diagnostic capability of the 3TP method. Color alone is insufficient for lesion characterization. Color intensity and hue provide important nuances. The wash-in enhancement rate is coded by color intensity. The wash-out phase—which reveals the change in enhancement between the second and third time points—is coded by color hue. More Time Points As described by the company, the pharmacokinetic element simulates vascular function as it analyzes contrast agent kinetics in tissue and clearance through kidney by spotlighting the MRI signal changes in cancer tissue. Further, the fTP image processing algorithm, now a core element in CAD Sciences’ products that characterize tissue and determine image colorization, generates numerical values for key physiological differentiators of cancer with a region of interest: vascular permeability and cellular density (or extracellular volume fraction). Simply put, the approach is based on identifying a certain profile of behavior of contrast agent as it comes in contact with both normal and cancerous tissue. fTP facilitates measuring and classifying the underlying tissue. “Instead of measuring the intensity of the contrast as it goes in and out of the lesion, fTP calculates these physiologic parameters, which allow us to discern the different biological processes taking place in tumors,” adds Wyszomierski. “With the permeability and extracellular volume fraction characteristic numbers for any given lesion, a radiologist can, for the first time, characterize a lesion, not just based on the general wash-in/wash-out principle, but on a very specific numerically driven. This may help the radiologist have a much more efficient communication with the referring physician.” Richard W. Reitherman, MD, PhD, of Orange Coast Memorial Breast Center in Fountain Valley, Calif., who has been using CAD Sciences’ technology, explains from a clinical perspective: “With three time points, users try to look at vascularity. But events in the blood vessel can happen at variable time intervals—after a minute, 90 seconds, or in some cases, two minutes. So the user chooses the second and third time points arbitrarily. With fTP analysis, the user is provided all of the data points, so they don’t miss inflections of curves, among other things.” Part of what happens with the inflow of contrast in breast lesions is that the user looks for the wash-in/wash-out time course of the injected agent, he continues. “But if you arbitrarily choose the wrong time point to find your peak, you may not locate it,” he says. “fTP provides a continuing curve, rather than just three curves on the graph, which offers more information, as it allows for the calculation of the permeability and extracellular fluid volume parameters.” Enhanced Patient Care He points to a specific, multifocal case involving a patient undergoing chemotherapy, where the first lesion responded but the second did not—information fTP can provide to radiologists and patients within 72 to 96 hours, according to Wyszomierski. “As a result of that information, the patient, in discussions with her oncologist and radiologist, could make an informed decision as to whether or not she should change the therapy treatment so that the second lesion would respond,” he relates. “She did and it was a successful decision, as the second lesion responded. That’s something you really can’t do that with a simple wash-in/wash-out concept.” According to CAD Sciences, there is a considerable body of published research literature demonstrating the usefulness of vascular permeability in monitoring response to therapy, which in some cases can be measured before and shortly after the start of therapy as early as 72 to 96 hours. fTP technology allows the user not only to visualize lesion characteristics in images but also to generate and compare important biological statistics—permeability and extracellular volume. That is something that is not possible with a simple wash-in/wash-out concept. It is easy to infer the strong ramifications that fTP technology can have in the future on the radiologist’s ability to confidently advise the referring physician in his decision to reaffirm or adjust therapy treatment going forward. Reitherman indicates that, for breast cancer, the patients best served by fTP fall into three categories: women newly diagnosed who need to know the extent of the disease and the most effective course of action, women already treated for cancer who need follow-up to watch for recurrence, and women with no known disease who seek additional screening. To illustrate the value of the related TRM, he offers an example involving inflammatory breast cancer, which presents in a rapidly progressive form, usually involving most of the breast. “In some cases, with fTP, we can show that chemotherapy produces dramatic results within a few days after the first treatment,” he says. “This is important to the oncologist, who can ascertain the effectiveness of a regimen, and to a patient, who has demonstrated poor prognosis in a rampant clinical course and now has information that her therapy actually helps.” Moreover, as Wyszomierski points out, fTP and its response-monitoring capabilities aid radiologists in their professional relationships with oncologists. “In cancer treatment, radiologists usually work with oncologists, who are, by definition and training, numerically driven. Conversely, radiologists are image-based. As such, radiologists have had trouble communicating to an oncologist about how a patient is responding. fTP enables radiologists, for the first time, to communicate effectively in terms that the oncologist is accustomed to regarding the effectiveness of a particular treatment plan,” he explains. “The fTP method is the only one that allows an MRI analysis to provide a radiologist with a numerical analysis of the characteristics of any given lesion.” He adds that fTP results in more advanced protocols: “Not only for the breast but a variety of organs such as the prostate, lung, and liver and on different modalities—computed tomography, as opposed to just MRI.” Innovative Solutions CAD Sciences also offers a TRM module that indicates a patient’s response to treatments such as chemotherapy and radiation and hormonal therapies, as it facilitates monitoring, over time, of vascular permeability, which is considered a key early physiological indicator of cancer. The measurements are made by Server 2.4’s advanced fTP algorithm and can help direct patients to the most effective therapies. “TRM allows us to look at details of the tumor before treatment and then, as we follow treatment over the course of time, to see the tumor cell responses to the therapy,” says Reitherman. “This allows us to predict further response, or no response, or perhaps the need to change the chemotherapeutic regimen. It’s an important feature for a dedicated breast practice, as radiologists see a number of patients who have neoadjuvant chemotherapy. It is critically important for the oncologist to have information about how the tumor is responding.” The module is versatile, says Wyszomierski. “It enables utilization of the fTP beyond a single organ, with the prostate being the first one with proven success after the breast,” he comments. For prostate cancer treatment, earlier this year, CAD Sciences introduced ProStream, a prostate MRI software product that also utilizes the fTP algorithm. ProStream resolves a significant dilemma: Contrast-enhanced prostate MR imaging requires a fast, dynamic imaging protocol because of quick contrast uptake in cancerous prostate tissue; at the same time, diagnosis requires high spatial resolution and good signal-to-noise ratios. The fTP algorithm allows for simultaneous high temporal and high spatial resolution in dynamic contrast-enhanced series. Also, the product can eliminate endorectal coil usage, which improves workflow, as coil placement involves a time-consuming, pre-imaging step. This not only quickens the procedure and removes costs associated with coils; it could also make prostate MRI more attractive for small, busy radiology practices, thus increasing application. Potential Applications TRM, he reports, is making it possible for other cancers to be analyzed for potential successful use with the fTP method. “Research is being carried out right now at institutes such as Duke University and Massachusetts General Hospital.” — Dan Harvey is a freelance writer based in
Wilmington, Del., and a frequent contributor to Radiology Today.
|
October
23, 2006
|
3801 Schuylkill Rd • Spring City, PA 19475 Publishers of Radiology Today All rights reserved. |