|
Giving cytotechs
a breather on
manual Pap test screening
FocalPoint GS system
classification model
ThinPrep Imaging System
October 2002
Karen Southwick
Cytotechnologists could be look ing at a far less fatiguing future.
Relieving the back-breaking tedium of manual Pap test screening is the aim of
new technologies being developed by two manufacturers of liquid-based Pap tests,
Cytyc Corp. and TriPath Imaging. The technologies, which consist of a computer,
analytical software, and motorized microscopes, direct the cytotechnologist
to certain locations, or fields of view, on the Pap slides that are more likely
to contain abnormal cells. This so-called location-guided screening, or LGS,
allows cytotechnologists to look at no more than two dozen fields of view, rather
than hundreds per slide, boosting productivity and reducing fatigue.
Neither Cytyc nor TriPath has LGS systems approved yet by the Food and Drug
Administration for use in the United States. TriPath’s FocalPoint Slide Profiler,
approved for primary screening, can triage slides into “normal” and
“possibly abnormal” categories, eliminating up to 25 percent of the
slides cytotechnologists must scan manually. More than 100 FocalPoint systems
are in place in the U.S. and Canada and in five European and six Asian-Pacific
countries. TriPath markets its GS system, with location-guided screening, outside
the U.S.
Clinical trials of both systems are underway in the United States, with FDA
approval and market launch expected sometime next year. Experts agree that LGS
could be not only an important step forward in cervical cancer screening, but
also a way to alleviate the chronic cytotechnologist shortage. However, early
users of the technology caution that, with reimbursement for Pap tests generally
low, the new systems will have to be priced cost-effectively to be widely accepted.
“This technology allows the cytotechnologist to spend less time looking
at cells on ThinPrep slides that are not of clinical interest,” says James
Linder, MD, chief medical officer of Cytyc and professor of pathology and microbiology
at the University of Nebraska Medical Center in Omaha. By weeding out obviously
normal cells, LGS screening lets cytotechnologists focus more. “The CTs
can accomplish more in a better working environment, using their skills to determine
whether a slide is normal or abnormal,” Dr. Linder adds.
The two products rely on different complex algorithms to zone in on fields of
view that may contain abnormal cells. “The choice of the object of interest
is based on the size of the cell and the amount of DNA inside the nucleus,”
says Dr. Linder. For example, “a cell with a large, dark nucleus and scant
cytoplasm is of greater interest than a cell with a small nucleus and abundant
cytoplasm.” TriPath’s approach uses complex algorithms to analyze cell
morphology, looking at the cell characteristics, including the size and shape
and optical density of the nucleus, says CAP_Cytopathology Committee member
David C. Wilbur, MD, who did preclinical work on the system.
Both systems are designed to work with liquid-based cytology (Cytyc’s with its
ThinPrep test and TriPath’s with its SurePath test). In addition, TriPath’s
system will work with conventional Pap smears.
The Cleveland Clinic Foundation is one of four sites that did
clinical trials on Cytyc’s ThinPrep Imaging System. Last year, the laboratory
used the system to look at about 2,700 Pap test samples and was enthusiastic
about the results, says Charles V. Biscotti, MD, staff pathologist.
“The analyzer identifies the fields that should be checked—the fields
with the most abnormal areas,” he says. When the cytotechnologist views
the slide under a microscope, the system will move the slide to enable him or
her to look at just the 22 fields of view designated as potentially abnormal.
“As long as the technologist decides there’s nothing worrisome, they sign
off on that case,” says Dr. Biscotti. “If there’s something worrisome,
the technologist will look at the whole slide manually.“
He likes the imaging system because it doesn’t attempt to replace human expertise,
but merely to use that expertise where it is most effective. “It doesn’t
replace cytotechnologists,” Dr. Biscotti notes. “It makes them more
efficient.” Similarly, the machine “doesn’t render any interpretation.
It just selects fields for the technologist to view.“
The system fits well into the laboratory’s workflow, reading the slides overnight
and having them ready in the morning for review by cytotechnologists. And the
cytotechnologists report that ergonomically, the ThinPrep Imaging System is
well designed because it is user-configurable. They control how they view the
selected objects of interest.
Cleveland Clinic cytotechnologist Dawn O’Brien says it was easy to use the keypad
to move among the selected fields. The device also will mark any group of cells
the cytotechnologist decides is atypical. “Then it would go back and dot
the field” for review by the pathologist, O’Brien says. If after reviewing
the 22 fields of view, the cytotechnologist decides there’s nothing abnormal,
“we didn’t rescreen the slide. If we think there’s something atypical,
we would rescreen the whole slide.“
Since a majority of Pap test slides are normal, using the device “saves
us a lot of time,” O’Brien says. “At the end of the day I had done
a lot more slides but I was less tired and it was much easier on my eyes.”
She estimates that in a typical day, she can manually screen between 70 and
80 slides. With the ThinPrep Imaging System, “I was looking at 115 to 130
slides.” She would like to use the system on a permanent basis, she says.
“It definitely makes my job easier.“
Although Cytyc has not released clinical results, Dr. Biscotti says he looked
at a sample of the 2,700 slides and determined there was about an 80 percent
increase in the numbers of slides screened per cytotechnologist. He wants to
see the clinical trial data to confirm his impression that the system’s accuracy
is at least comparable to or better than that of manual screening.
Assuming the accuracy data are conclusive, once the FDA gives its approval,
“I would push for its use here,” Dr. Biscotti says. “I like the
technology.” Of course, “I have to look at the economics and see whether
the improved efficiency will decrease the labor cost per unit enough to justify
the cost.” Deciding whether to buy LGS devices may depend on the pricing
set by the manufacturers, the lab’s Pap test volume, and whether it has and
can retain an adequate number of cytotechnologists, he suggests.
Two other experts who have seen demonstrations of the ThinPrep Imaging System
also give it high marks. “I found it easy to use, the learning curve is
very shallow, and ergonomically it’s designed well,” says Ralph M. Richart,
MD, vice chairman of anatomic pathology at Columbia University College of Physicians
and Surgeons in New York.
Dr. Richart says the system will fit easily into a lab’s workflow because it
can prescreen the Pap specimens unattended and have them ready for cytotechnologist
review. “The advantage is that it selects the most atypical cells and presents
those as a series of fields to the CT,” he says. “The data I’ve seen
suggest the machine will do as well or better than a CT doing a manual screen,
while doubling the throughput.“
Like Dr. Biscotti, Dr. Richart is relieved that the device combines technology
with human expertise. “A device and software is still not as good at interpreting
abnormal cells as a human,” he says. “This device finds an array of
cells that could be abnormal, but the human does the interpretation.“
Karen Allen, SCT (ASCP), of Heartland Pathology in Omaha, Neb., and president
of the American Society for Cytotechnology, has participated in two hands-on
demonstrations of the ThinPrep Imaging System. Besides the ease of looking at
the selected fields of view, she also liked that the machine will mark selected
fields that the cytotechnologist might want to make comments about. The system
can be linked to the lab network and the pathologist “will have the CT’s
comments as they relate to identified locations on the slide.“
Cytotechnologists now must look at every field and every slide “from corner
to corner” to spot abnormalities, Allen notes. “There’s a lot more
intensive mental decision-making with manual screening.” Allen sees four
advantages to the LGS system: less eyestrain; fewer repetitive stress injuries
in the arms, neck, and back; higher productivity; and better interpretation
accuracy. “The machine may be able to pick up some things a human might
miss,” she says.
LGS should also help stretch the supply of cytotechnologists by making them
more productive and their jobs more attractive. Nationwide, there’s more than
a 20 percent vacancy rate for cytotech jobs, she says. “CTs are retiring
and we’re not producing as many in the schools. The timing for these systems
is great.“
Allen’s interest in using the system at her own lab, where she is president
and owner, is strong. But “I wouldn’t be able to determine whether or not
to purchase it without knowing the price the manufacturer will set,” she
says. Large commercial labs will see the greatest advantage, she says. “The
more tests you put through, the more economical it will be.“
Dr. Wilbur did the preclinical feasibility trial work on
TriPath’s FocalPoint System at ViaHealth in Rochester, NY. He is now director
of cytopathology at Massachusetts General Hospital in Boston and associate professor
of pathology at Harvard Medical School.
In his research comparing the FocalPoint technology with manual screening, using
1,275 slides, LGS identified more abnormal cases at all interpretive levels
and classified abnormalities more specifically, Dr. Wilbur says. (The article,
“Location-guided screening of liquid-based cervical cytology specimens,”
is in the September issue of American Journal of Clinical Pathology.
2002; 118: 399–407.)
The overall sensitivity of appropriate triage to pathologist review for high-grade
squamous intraepithelial lesions, or HSIL, was 98.4 percent for LGS versus 91.1
percent for manual screening by cytotechnologists. The appropriate triage for
all abnormal cases was 92.1 percent for LGS and 87.9 percent for manual screening.
“Our study was intended to establish the operating characteristics of the
protocol that the clinical trial will use,” Dr. Wilbur says. For example,
“how many fields of view do you need” to demonstrate efficacy with
LGS?
The FocalPoint System, already approved to triage slides for further manual
review or no need for further manual review, can weed out up to 25 percent of
slides as not requiring cytotechnologist screening. “The optimal working
rate is 22 to 23 percent,” Dr. Wilbur says. For most labs starting with
FocalPoint, “it will likely be around 20 percent, but will move closer
to the maximum level of 25 percent with optimization of slide preparation and
experience.“
The FocalPoint System automatically reviews slides using a high-speed video
microscope. On an initial low-magnification scan, it identifies the areas containing
cells and then does a high-magnification scan, identifying fields of view for
the cytotechnologist to review (in the LGS system only). In the FDA-approved
current version, it ranks slides based on the probability of their containing
abnormality, then divides the slides into those not needing and those needing
manual review. For those requiring a manual review, FocalPoint also indicates
those having the highest probability of abnormality.
In addition, after manual screening, it suggests which slides should undergo
a quality control check. “If you screen this slide and decide it’s negative,
the machine suggests you QC it,” says Dr. Wilbur. “Instead of a random
QC process, it’s a 15 percent directed QC with only the highest-scoring slides.“
Of the three to four papers published on LGS, “all have found improvements
in performance, both in reducing false-positives and false-negatives,”
he says.
Dr. Wilbur predicts manufacturers will offer a range of pricing. “Many
labs will purchase the machines on a per-click basis,” meaning there would
be a fee per slide. Labs that process at least 30,000 to 40,000 slides annually
are in the best position to make economical use of this technology, he adds.
Massachusetts General Hospital is in the process of converting to the FocalPoint
technology, which it will use in its current FDA version until the FDA approves
the LGS enhancement. “We’re using it to improve our cytotechnologists’
productivity,” Dr. Wilbur says. While he doesn’t expect to make money with
the machines, “hopefully we won’t lose any and we’re improving both quality
of service as well as productivity.”
With the cytotechnologist shortage projected to worsen, one of the recommendations
of a current American Society of Cytopathology task force dealing with the shortage
is to “promote acceptance of this kind of technology,” he says. “We’ve
got increasing volumes of Pap smears and decreasing sources of screening capability.
We need this technology.“
Heinrich Neumann, MD, FIAC, a pathologist with the Institut fur Pathologie in
Nordhorn, Germany, has been evaluating the FocalPoint System for about two years.
His laboratory assisted TriPath in refining the software to accommodate the
change from screening an entire conventional slide to homing in on the defined
circular area and to tolerate a few cells outside this circle. “Initially
we had problems with the screener’s performance because it rejected too many
slides,” he says. “We improved our slide quality, especially with
coverslipping, and in parallel TriPath developed new software.“
From June to August of this year, FocalPoint Slide Profiler screened 11,600
SurePath slides at the Institut. Now only 0.6 percent of slides had to be rerun
because of technical problems, and 2.6 percent of all slides could not be analyzed
by the instrument. The instrument did a diagnostic evaluation of the remaining
(qualified) 96.8 percent of the total slides. Dr. Neumann set the parameters
so that 25 percent of qualified slides would be signed off as needing no further
review. In reality, about 24.4 percent of qualified or 23.7 percent of all slides
could be signed off in that way by the instrument. The mean time the FocalPoint
Slide Profiler took to analyze these 11,600 slides—327 seconds per slide—would
suggest an annual screening capacity of more than 90,000 slides, Dr. Neumann
estimates.
For each field of view, “the instrument creates a value of abnormality,”
Dr. Neumann says. He uses 15 fields of view selected by the instrument for review,
excluding the 25 percent the machine has already triaged. “You can very
easily click through the 15 fields and get an impression if this slide is abnormal
or not. A normal microscope is used, which is very comfortable for the microscopist.
You’re able to use a foot-switch in addition to the mouse to drive the system.“
What the FocalPoint system does not do as accurately is recognize normal endocervical
cells, “so if you want precise EC-negative rates, you must still manually
review all the slides,” Dr. Neumann says. Even if the machine has reported
“no further review,” his lab does a quick manual screen, he says.
In the 1 1/2 years he’s run the system, “we found a handful of LSIL [low-grade
squamous intraepithelial lesions] and ASC-US [atypical squamous cells, undetermined
significance] that the machine had designated ’no further review,’” Dr.
Neumann says. So far, no high-grade squamous intraepithelial lesions have been
designated as requiring no further review. He’s beginning an additional study
now to compare LGS with manual screening. “We will be collecting these
data in the next few months,” Dr. Neumann says.
In Germany, where reimbursement for Pap tests is very low, “you’ll lose
money on this technology if you don’t take it seriously and commit yourself
to using it,” he says. Combined with liquid cytology, it does increase
productivity. Even with endocervical scans, cytotechnologists can read up to
100 slides per day compared with 75 to 80 SurePath slides using manual screening.
“Buying this technology depends on how much volume you’re doing and how
bad the shortage of cytotechs is,” Dr. Neumann says. “If you’re really
short of manpower, it might be a good idea to purchase the machine.” If
not, he notes, switching to liquid-based cytology could improve quality and
productivity without the additional cost of the automated screener.
Cytyc and TriPath are hopeful their products will be launched
next year. Cytyc’s Dr. Linder says its trial results have been submitted to
the FDA. “We have not released trial data publicly,” he says. “The
design of the trial was intended to demonstrate both improved lab throughput
and the potential for improved diagnostic accuracy. It’s fair to say we’re pleased
with the trial results.“
He expects an FDA advisory panel to meet this fall or early next year on the
Cytyc submission, which is for premarket approval involving cancer screening.
“We expect approval would come shortly thereafter,” Dr. Linder says.
Trials are underway in five countries in Europe, and Dr. Linder expects the
ThinPrep Imaging System to be introduced there this winter. “In Europe
the shortage of CTs is even worse, and there’s keen interest in the product,”
he says.
As for pricing, “we designed the system not anticipating any additional
reimbursement,” Dr. Linder says. Rather, “we believe our pricing must
allow the lab to achieve cost savings.” Because the Cytyc system is designed
around the ThinPrep liquid cytology, it should facilitate ThinPrep use, allowing
the company to price competitively. Dr. Linder estimates that the ThinPrep Imaging
System, consisting of an imager, computer, and three review microscopes, can
handle more than 80,000 Pap tests a year.
TriPath is now initiating its LGS clinical trial in the U.S., and the trial
is expected to last from four to six months, says Ray Swanson, senior vice president
for commercial operations. TriPath intends to submit a premarket approval application
for the FocalPoint system in 2003.
“We haven’t announced an LGS-specific pricing model yet,” he adds,
but, like Cytyc, TriPath expects LGS to drive sales of its liquid cytology.
Swanson says TriPath expects to negotiate acceptance of the LGS system, as it
does today for the FocalPoint primary screener. “We will negotiate with
labs to offer the most flexible pricing, including a capital lease or reagent
rental. We want to find the right value proposition for the lab,” he says.
That proposition will be based on several variables: the system’s impact on
the lab’s screening accuracy and economics, and on the lab’s reimbursement,
since conventional slides screened by the FocalPoint are reimbursed at a premium
to manual processing, Swanson says. “We expect labs will be able to use
this product to differentiate themselves,” he proclaims.
While the FocalPoint GS system in Japan uses 15 fields of view, “we anticipate
claims for 10 to 15 fields of view in the U.S.,” Swanson says. The existing
FocalPoint can process up to 75,000 Pap tests a year, and “we expect upwards
of 95,000 or more for the GS system,” he says.
Cytyc and TriPath will be extremely competitive in the United States, Swanson
acknowledges, where the Pap test market is moving toward liquid-based cytology.
In other countries, however, Swanson says the FocalPoint GS system may have
a leg up because it can handle traditional Pap smears. “Outside the U.S.
we feel we have an advantage because reimbursement is poor, so they’re not doing
liquid cytology. They’re primarily doing conventional cytology,” he says.
Karen Southwick is a writer in San Francisco
|
|
|
