IVD Technology Magazine | IVDT Article Index
Originally published March, 1997
IVD Technology News
Automated Pap systems experimentally check sputum
The term "spitting image" is taking on new meaning. Two fully automated Pap tests--approved less than two years ago to root out false-negative cervical smears--now are being used on sputum samples, too.
This new application is not likely to be submitted for FDA clearance anytime soon. In the United States, these automated cancer-screening devices are approved solely as a back-up check for the Pap test, limited strictly to the realm of quality control in the assessment of cervical cells.
But representatives for both tests, Papnet and AutoPap QC, say they are trying out the technology in other organ systems, most notably the lungs, with an eye toward the overseas market.
"We think the market is already there," says James S. Lee, PhD, vice president and chief scientist at NeoPath, Inc. (Redmond, WA), which produces the AutoPap. He points out that in Japan and Germany, sputum smears are read like cervical samples: by eye. A computer-assisted method is appealing, in terms of both cost and accuracy.
"The technology may have other applications as well," says Lee. "Esophageal cancer, melanoma, breast cancer...all these are possibilities."
Neuromedical Systems (Suffern, NY), which developed the Papnet system, is already busy researching the potential of its equipment for evaluating sputum smears to diagnose lung cancer. At St. Mary's Hospital (Junction City, CO) and at the Albert Einstein College of Medicine in New York, the Papnet system identified 30 of 31 positive slides--with no false positives. One physician re- searcher, Gene Saccomanno, MD, noted that trying to evaluate these smears by sight, as in Pap smear screening, "is tedious and time-consuming." In a news release on the findings, he suggested that automated screening may reduce the need for this labor-intensive process.
Both Papnet and AutoPap use high-resolution digital scanners, with a computer that peruses the samples for suspicious variations. Images with cells that appear abnormal are flagged for further human review. How helpful is the technology in rescreening? Developers say it can cut the false- negative rate practically to zero.
Thanks to heavy coverage in the media, many women now request this "computer test." As Nancy Snyderman, MD, proclaimed on a segment of Good Morning America this past winter: "It doesn't make sense" to rely on human eyes when the equivalent of
a cellular "spell checker" is available.
But not all physicians endorse the technology so wholeheartedly. The American College of Obstetricians and Gynecologists (ACOG), while pronouncing the technology promising, wants further testing. "I tell my patients that these techniques show promise for increasing accuracy, but that current quality control methods have improved accuracy," Kenneth Noller, MD, noted in a recent ACOG newsletter.
Noller, who is professor and chair of obstetrics and gynecology at the University of Massachusetts Medical School in Worcester, added that "rescreening is not 100% accurate and never will be." A similar view prevailed last fall at a conference cosponsored by the American Medical Association, "Examining Errors in Health Care" (Rancho Mirage, CA, October 1315, 1996).
David Kaminsky, MD, who heads the department of pathology at the Eisenhower Medical Center (Rancho Mirage), asserted that "the consumer has been cultivated to believe this is a perfect test." In fact, the time-tested Pap has had a false-negative rate of about 5% for years. By itself, neither new automated system "has been proven to reduce that error rate," Kaminsky added.
However, many precancerous lesions no longer slip through the screening process, according to published journal reports. Last year, in two multicenter studies, NeoPath investigators compared conventional and automated evaluations of 2584 slides to determine whether the abnormal cells would be categorized identically by the two methods.
Not only did the classifications prove concordant, but AutoPap also picked up subtle precancerous signs that otherwise would have been missed (Acta Cytologica, 40:4552, 1996). In a study of 101 slides previously determined by conventional pathologic evaluation to demonstrate abnormalities, Neuromedical Systems' Papnet results concurred, to a large extent, with the original reviews. But Papnet-directed rescreening also led reviewers to reclassify 35 cases, about a third of which were upgraded to a higher risk category (Am J Clin Pathol, 105:711 718, 1996).
But, according to Mitchell Ryan, MD, the senior author of the Papnet study, "this needs to be confirmed with more cases." Ryan, an instructor in pathology at Washington University Medical Center (St. Louis), believes that the next time around this kind of study should be done "with larger numbers of samples."
And here's another rub, says David Wilbur, MD, director of cytopathology at the University of Rochester in New York and a consultant to NeoPath. "If you ask 'Is it cost effective?' the answer is 'It can be, in certain situations.'
"The biggest promise for cost-effectiveness comes with primary screening," Wilbur continues. NeoPath hopes to find out by the end of the year what the chances are that FDA will approve its PMA request for use of the device in primary screening.
Lee envisions a synergistic role for the two automated systems. Theoretically, he posits, "you could use one system in a primary screening process, and do quality control with another system. You could use two very different computer-assisted screening technologies back to back."--A.S.
