The high proportion of false negatives associated with Pap smears spurred
the development of more effective collection and diagnostic techniques
for cervical cells. The Pap smear, a universal standard in the detection
of cervical cancer, has recently been refined by two new approaches that
greatly reduce the incidence of false negatives.
At the annual meeting of the Society of Gynecologic Oncologists, study
data was presented that showed that cervical cancer and precancer detection
can be enhanced by the use of neural network (artificial intelligence)
computer technology. More than 16,000 smears, reprocessed by the computer-based
PAPNET test, found many pre-cancerous abnormalities that had been missed
by conventional screening. These slides were originally reported as negative
despite presenting with an abnormality. PAPNET testing was shown to identify
10 times more false-negative smears than would have been detected by manually
rescreening an equal number of smears.
The study was conducted by Laurie J. Mango, MD, of Neuromedical Systems,
Inc. developer of the PAPNET test, and Gary L. Goldberg, MD and Irwin R.
Merkatz, MD, both of the Albert Einstein College of Medicine. "Negative"
smears were reviewed and reclassified across four categories: ASCUS/AGUS
(atypical), LGSIL (low-grade squamous intraepithelial lesion), HGSIL (high
grade), and carcinoma. Three academic laboratory sites provided the archived
smears for review.
Mead Johnson and company recently agreed to copromote another refinement
to the Pap smear. Cytyc corporation launched its ThinPrep Pap Test in January,
1997. In clinical trials , the test has proven to be significantly more
effective than the conventional smear for the detection of low-grade and
more severe lesions and to markedly improve specimen quality. Mead Johnson
signed on to distribute and market the test to obstetricians and gynecologists
throughout the United States. Currently, over 400 laboratories in the United
States employ the ThinPrep system for use in the diagnosis of various cancers.
The ThinPrep System consists of the ThinPrep 2000 Processor and related
reagents, filters and other supplies.