| |||||||||||||||
|
Home
|
For other articles and previous issues click here. May 8, 2006
The CAD for Lung Cancer — Still Progressing Lung CAD—computer-aided detection applied to lung cancer diagnosis—first entered the collective consciousness of the general medical imaging community around the time of RSNA 2004. That was the year that witnessed the introduction of Riverain Medical, LLC’s RapidScreen technology, which became the only FDA-approved CAD system for chest x-rays. In the ensuing year and a half, interest in lung CAD took a definite upward swing. “At RSNA 2004, people were curious but not all that knowledgeable about CAD, and somewhat skeptical. At RSNA 2005, interest became huge,” says Terry Chang, director of CT marketing for the Sunnyvale, Calif.-based R2 Technology, Inc. Chang adds that subsequent sales of R2 Technology’s ImageChecker CT Lung CAD system reflected the surge. Scott Pohlman, director of clinical science for Philips Medical Systems, voices similar observations. “At RSNA 2005, increased interest was evident in the greater number of CAD presentations, as well as the attendance at these presentations,” he recalls. He experienced more direct evidence on the 2005 RSNA showroom floor, where he and his company fielded many questions from customers about CAD. Philips is developing its own CAD lung system for CT. Lung CAD software—developed for both chest radiographs and CT—provides a virtual second opinion from the collaboration between radiologist and computer. It seeks to address the No. 1 cause of cancer deaths in the United States. Lung cancer is among the hardest forms of cancer to effectively treat because it is rarely diagnosed early. The best hope for survival beyond five years is early diagnosis, which can be hard to accomplish. According to the American Cancer Society, 160,000 people die each year from the disease and, currently, only approximately 15% of patients survive five years after diagnosis. Improving
Outcomes? The April 2005 issue of Radiology reported research from the Mayo Clinic reinforcing earlier conclusions about lung cancer screening that positive low-dose CT (LDCT) exams don’t correlate with better outcomes. Too often, Mayo concludes, LDCT lung scans identify benign anomalies as suspicious growths, leading to unneeded expense, therapy, and surgery. For the most aggressive cancers, in most cases, early detection, whether by LDCT or other means, simply makes no difference in either treatment planning or outcomes, according to Mayo clinic research. The challenge remains in defining which patients should be screened and when they should be screened to improve outcomes. Formal guidelines from the National Cancer Institute, the American College of Chest Physicians, and the American Medical Association caution that available studies on spiral CT scanning for lung cancer cannot yet positively demonstrate a correlation between regular screening and reduced mortality. Still, interest in CAD is growing, driven in many people’s views by the following four factors: • awareness of new studies that indicate CAD’s viability in detecting lung cancer earlier; • increased usage of multislice detector computed tomography (MDCT); • advancements that increase sensitivity and specificity and reduce false positives; and • promising news on the reimbursement front. “The marketplace already perceives CAD as a fast-track, emerging technology. People know about the promise and opportunity,” says Sam Finkelstein, president of Riverain Medical, LLC, the Miamisburg, Ohio-based company that has pioneered lung CAD technology with chest x-rays. “Now, they’re looking for certain product performance levels that assure them two things: that the product gives them a clinical confirmation tool for early lung cancer detection and that it has a positive impact on their workflow.” Finkelstein says people generally accept that CAD can improve survival rate by helping radiologists identify early-stage nodules, which have been very difficult to detect. The RSNA 2005 meeting included a greater number of presentations involving lung CAD. One of the most illuminating (“Results from a Commercial CAD Program in Cases of Missed Lung Cancer, Visible in Retrospect on Chest Radiography”) concerned a study conducted by researchers from the department of radiology at the University of Chicago (Heber MacMahon, MD; Feng Li, MD; Kunio Doi, PhD; and Roger Engelmann, MS). For their study, the researchers, who used Riverain’s RapidScreen chest x-ray CAD system, set out to determine the effectiveness of CAD for detecting lung cancers on chest radiographs that had been missed at the time of interpretation. MacMahon and colleagues analyzed relevant radiographs and reports from the University of Chicago Hospital’s cancer registry for patients diagnosed with lung cancer since January 2001. Working with 831 case records, they identified 49 prior radiographs for patients where nodular primary lung cancer was noted only in retrospect. From the 49 radiographs, 38 cases were rated as probably “actionable.” Researchers concluded that the CAD technology available could increase sensitivity for detecting subtle lung cancer on chest radiographs. And that technology is improving. “We’ve advanced our algorithm and have several additional premarket [approval (PMA)] applications in the pipeline for submission to the FDA,” Finkelstein says. “The improved algorithms will provide greater sensitivity and specificity.” CAD
and CT During the study, chest CT exams performed on 100 patients were initially reported as normal at clinical double reading. Nevertheless, CAD then detected significant lung lesions in 33% of those. In all, 53 significant lesions were detected (1.6 lesions per case). Software
Evolving Temporal comparison automatically tracks lung nodule changes over time. Chang considers the capability very crucial in diagnosing and treating lung cancer. “CAD has always been known as a detection tool, but detection is really only half the battle,” he points out. “Once you’ve found the nodule, then you have to do something about it.” Typically, a detected nodule leads to a biopsy. But lung nodules are hard to biopsy, and they may be benign. “You need to track its growth over time, to determine malignancy,” says Chang. “Normally, nodule growth is compared in two dimensions. But nodules grow in three dimensions. They’re like a balloon.” Temporal comparison tool, like R2’s AutoPoint, tracks changes volumetrically over time, which helps reduce reader variability. “When a computer measures a nodule, it does it the same way over time,” says Chang. In addition, such volumetric tracking adds a high degree of efficiency to MDCT studies, which can involve as many as 1,000 slices. It takes a lot of time to hunt down and match nodules in current and prior studies. Also, the AutoPoint uses an automatic registration feature that anatomically matches nodules. “Matching by anatomical reference points is far superior than matching by slice number,” says Chang. “When you match by slice number, you’re not taking into account the variability in a patient’s respiration. This eliminates the variability, because a patient’s anatomy doesn’t change over time.” Already, the company has an upgraded 2.1 version in development. “This third iteration, which will come out later this year, has improved algorithms and will make CAD much more efficient in a physician’s workflow,” says Chang. Philips will soon offer lung CAD technology that
also offers temporal comparison capability. Currently, the Philips
Lung Nodule CAD is available only outside North America. It is not
approved in the United States and Canada. Already, lung CAD vendors recognize the importance of PACS connectivity and are expanding their products appropriately. “We’re developing a universal connectivity with different PACS,” reveals Finkelstein. “Users will be able to deploy our product without having to modify their PACS or work with different PACS manufacturers to accommodate our product, which is a common challenge in the industry.” Similarly, R2’s ImageChecker CT Lung CAD system features expanded PACS integration for increased workflow efficiency. It enables CAD results to be sent to a facility’s PACS, using standard formats for enterprisewide accessibility and enhanced workflow. “Radiologists spend a lot of their day on PACS. It’s where they want to do their reads across multiple modalities,” says Chang. “Without this ability, a physician would have to go to a separate work station, which is very disruptive of their workflow. We’ve now automated the ability to send the CAD results and measurements to any PACS using a DICOM format. We can’t forget that CAD has to work in the workflow of the radiologist. It’s not supposed to make their work harder.” Reimbursement is the second major issue. “It’s what changes a technology from being an interesting new application looked at by a few early adopters to being something that could have widespread application,” says Pohlman. Earlier this year, the American Medical Association issued promising news for CAD applied to chest x-rays when it announced a Category III CPT code (Code 0152T) to cover chest x-ray CAD. The code, which became effective January 1, states: “0152T Computer-aided detection (computer algorithm analysis of digital image data for lesion detection) with further physician review for interpretation, with or without digitization of film radiographic images; chest radiograph(s) (List separately in addition to code for primary procedure) (Use 0152T in conjunction with 71010, 71020, 72021, 71022, and 71030).” Still, the code doesn’t guarantee reimbursement. Rather, creation of the code indicates that payors understand the benefit of CAD analysis of x-rays, and it will make it easier for facilities to obtain financial reimbursement for FDA-approved CAD technology. The code is already having a positive impact: Riverain reported that inquiries about its technology substantially increased. Lung CAD for CT exams is not yet reimbursable. But Chang is optimistic. “We believe that there are existing codes that our lung CAD could use,” he says. “We feel that we are in a very good position to get reimbursement. We have already shown the clinical efficacy through the PMA process of the FDA. So, at this point, it is just trying to convince the payors.” Toward
the Future —
Dan Harvey is a freelance writer based in Wilmington, Del., and
a frequent contributor to Radiology Today.
|
 |
|
3801 Schuylkill Rd • Spring City, PA 19475 Publishers of Radiology Today All rights reserved. |