![]()
13
| 4 | 2006
Klinisk Biokemi i Norden
(Fortsætter side 14)
(n = 4) of the pathological samples were not identi-
fied by the digital instrument. In two cases this was
due to minor differences in distribution and in one
case it was due to the presence of a single blast cell.
In one case the distribution was markedly different
but in this sample the results from the DiffMaster
correlated closely with those from the cell counter.
The clinical specificity was lower (82 %). This
was due to the fact that the DiffMaster identified
more immature cells than the manual reference
method. Imprecision levels for the two methods
were similar.
Practical experiences and considerations
The DiffMaster was introduced into the clinical
routine more than three years ago and is now used
in the laboratory for > 90 % of the slides requiring
microscopy review.
The throughput varies according to the count-
ing requirements of the specimens being handled.
The number of cells counted on each slide can be
adjusted; for 200 cell counts per slide the through-
put is 12 – 15 slides per hour; for 400 cell counts
per slide, the throughput is 6 – 9 slides per hour.
This included analysis of red cell morphology.
The equipment has several advantages: It is very
easy to train the operators as the software can be
navigated intuitively. The greatest advantage and
what has made it appreciated among microscopists
is that, using the zoom facility, it gives an improved
view of the cells, especially we feel, that the imma-
ture characteristics of the cells are easier to appre-
ciate on the screen. Ergonomics are also improved,
although attention has to be given to amount of
time spent using the mouse. Standardisation has
improved in the laboratory and if a given test result
is questioned, it can be retrieved and re-evaluated.
Further advantages are summed up in table 2.
One special requirement that may cause problems
is the need for high quality smears. These digital
systems are not as robust in correcting for smearing
mistakes and smearing errors as the human eye. In
contrast when using an automated slide maker and
staining machine these problems have not been
encountered. We have also used a manual device
successfully for making blood smears (Hemaprep).
Since this first evaluation CellaVision AB has
introduced a new system to the market named
CellaVision ™ Octavia, which has improved software
and continuous slide feeding, with an initial load
of 96 slides.
TeleHematology
The DiffMaster instrument and the not automatic
LAFIA system (Sysmex), can be networked and be
located in different hospitals. Centralised reviewing
and case conferences are possible. If further evalu-
ation is needed, digital images can be sent out to
different laboratories for education and training of
the staff.
A project in collaboration with CellaVision AB
and the External Quality Assurance in Laboratory
medicine in Sweden (EQUALIS) ( Birgitta Swolin
and Gunnar Nordin) a programme for digital mor-
phological cell classification has been developed,
CellAtlas EQUATOR. Since two years this program
has been available for laboratories in Sweden. The
images of the cells are stored in a JPEG format, so
that they may be e-mailed within the organization
or to colleagues for comment and assessment. This
program offers a possibility to standardise the mor-
phological competence, nomenclature and as well
define terms.
The digital format offers new possibilities, such
as following how a patient responds to a particular
treatment. This possibility has, however, not yet
been explored in our labs. Finally training of the
personal has also improved, as the trainee now first
alone classifies the samples, and subsequently the
sample can be evaluated by the teacher.
Advantages of computerised
pattern recognition systems in
clinical haematology
• Standardisation
• Improved workflow
• Reduced overall turn-a-round times
• Zoom facility improves view of the cells
• Easy to train the operators
• Improved ergonomics
• Online archive
• Documentation for patient record
Table 2
