34
Automation in Cervical Cytology
F ig . 3 4 .1 8 T h in P re p Im a g e r S y s te m re v ie w s c o p e s ta tio n . In a fashion
sim ilar to th e FocalP oint re vie w station, th e ThinP rep Im a g in g System
re vie w sta tio n receives co o rd in a te s fro m th e im ag e r and a u to m a tica lly
directs th e c y to lo g is t to h ig h -p ro b a b ility fields o f v ie w (FOVs) fo r rapid
e xam in a tio n . T w e n ty -tw o FOVs are gen e ra lly d e sig n ated o n each slide, and
if no a b n o rm a lity (or p o te n tia l a b n o rm a lity) is id e n tifie d o n th a t review , th e
slide can be d e sig n a te d as n egative w ith o u t fu rth e r m anual screening. If
an p o te n tia l a b n o rm a lity is id e n tifie d , c o m p le te slide re vie w is re quired to
ensure th a t th e m o st accurate classification ensues. The m icro sco p e is fu lly
a u to m a te d and is c o n tro lle d by a m ouse-like "p o d ” d evice seen on th e rig h t
side o f th e scope.
reviews the id en tifie d FOVs, w h ic h are relocated a uto m a tica lly
b y the use o f a m ouse-driven integrated ro b o tic stage. FOVs are
relocated geographically fro m one end o f the cellular circle to
the o th e r (as opposed to the FocalP oint hierarchic ran kin g proc-
ess). I f no a b n o rm a lity is detected in the 22-FO V review, the slide
m ay be signed o u t as N IL M w ith o u t fu rth e r m an ua l screening. I f
any p o ten tial a b n o rm a lity is id entified , the slide m ust go on fo r
a com plete m anual screening, w h ic h is perform ed o n the same
m icroscope in an autom ated fashion. A fte r a fu ll m anual screen-
ing, there is the o p tio n to a uto m a tica lly place in k dots on the
slide fo r fu rth e r non-d evice-aid ed slide review. A n image o f the
m icroscopic review fie ld w ith cell lo c a liza tio n reticle is show n in
Fig. 3 4 .1 9 . Fiducial m arkings (see Fig. 34.4B) present in the
T hinP re p slide ensure proper lo c a liza tio n o f the device-selected
FO V in the review microscope.
C linical
trials
o f
the
ThinP rep
Im aging
System
have
show n statistically equivalent sensitivity fo r the detection o f
H S IL and greater and LSIL and greater lesions and statistically
F ig . 3 4 .1 9 T h in P re p Im a g e r S y s te m re vie w scope m icro sco p ic fie ld o f
view . The v ie w in th e m icrosco p e h ig h lig h ts th e central area o f interest as
d e sig n a te d by th e ThinP rep Im a g in g System scanner. The
L-shaped reticle show s th e loca tio n w h e re th e a u to m a te d d o ttin g fu n c tio n
th a t is in h e re n t in th e ThinP rep Im a g in g System re vie w scope w ill place a d o t
a t th e request o f th e c y to lo g is t on c o m p le tio n o f th e re vie w process.
increased specificity fo r the detection o f H S IL and greater
lesions. In a d d itio n, statistically sup erior se n sitivity fo r the
detection o f ASCUS and greater lesions and statistically equiv-
alent specificity fo r the detection o f ASCUS and greater and
LSIL and greater lesions were show n in the clinical tria l data.45
P ro d u c tivity u tiliz in g th is process im p roved substantially, such
th a t U SFD A lab eling o f the device allow ed cytology screeners
to exam ine 200 F O V -o n ly cases d uring an 8-h w o rk period, as
opposed to the current C LIA -m andated m a x im u m o f 100 slides
per 8-h period. The process continues to im p le m e n t the cur-
re n t C LIA -m andated 10% rand om and targeted Q C rescreening
req uirem ent, s im ila r to fu lly m anual screening.
Post approval, a n um b er o f clinical studies have docu-
m ented field perform ance o f the device. In tw o large stud-
ies
com paring
im ager
perform ance
on
ap p roxim ately
80 0 00 -1 0 0 000 cases each w ith an histo ric al c on trol p op ula-
tio n o f s im ila r num bers o f m a n u a lly screened cases, the T h in -
Prep Im aging System id en tifie d m ore LSILs (8 and 37% ) and
m ore HSILs (13 and 4 2 % ).109,110 O ne o f the studies showed an
estim ated false-negative rate th a t was halved fo r th e studied
p o p ulatio ns w h en using the im aging device.110 A n o th e r study
showed s im ila r increases in SIL cases id en tifie d b u t also docu-
m ented a decrease in H P V p o s itiv ity in cases o f ASCUS id en tifie d
using the imager, ind icating th a t increases in detection o f this
equivocal category are m ost lik e ly due to overcalls o f reactive
find ing s.111 O ne im p o rta n t u tility gained by the autom ated
screening process is im p roved prod uctivity. A p ostim p le m e n ta -
tio n tim in g study showed th a t overall screening tim e using the
device was reduced b y 4 2% over m anual screening.112 Interest-
ingly, in th is study a n um b er o f im ager "false negative" cases
were id en tifie d by m anual screening. O n review o f each o f these
cases, all had ab norm al cells w ith in the device-identified FOVs,
ind icating th a t operator, and n o t device, error was responsi-
ble fo r the m issed cases. T he authors indicate th e necessity fo r
review o f the entire FO V and n o t ju st the central areas in order
1035
previous page 1019 ComprehensiveCytopathology 1104p 2008 read online next page 1021 ComprehensiveCytopathology 1104p 2008 read online Home Toggle text on/off