433

ORIGINAL ARTICLE

Digital slide and virtual microscopy based routine and

J Clin Pathol: first published as 10.1136/jcp.56.6.433 on 1 June 2003. Downloaded from http://jcp.bmj.com/ on 27 May 2019 by guest. Protected by copyright.
telepathology evaluation of routine gastrointestinal
biopsy specimens
B Molnar, L Berczi, C Diczhazy, A Tagscherer, S V Varga, B Szende, Z Tulassay
.............................................................................................................................

J Clin Pathol 2003;56:433–438

Aims: To evaluate a recently developed digital slide and virtual microscope system, and to compare
this method with optical microscopy on routine gastrointestinal biopsy specimens in both local and
remote access modes.
Methods: A fully computer controlled commercial microscope was used. The scanning program
included object detection, autofocus, and image compression algorithms. The overall hard disk space
for a gastric biopsy was between 30 and 50 MB and the scanning time was between 20 and 40 min-
See end of article for utes. Haematoxylin and eosin stained routine gastric (61) and colon (42) biopsy specimens were
authors’ affiliations
....................... selected, scanned, and evaluated by two specialists on an optical (OM) and virtual microscope (VM).
Results: The overall concordance of VM and OM with the consensus diagnosis was 95.1% and 97%,
Correspondence to: respectively. Clinically important concordance was 96.1% and 98% for VM and OM, respectively. The
Dr B Molnar, II Department
of Medicine, Semmelweis
two methods showed concordance in 92% of cases and clinically important concordance in 94.1% of
University, Szentkirályi St. cases. The reasons for discordance were image quality (one case), interpretation difference (three
46, 1088 Budapest, cases), and insufficient clinical information (three cases). Remote evaluation of the digital slides through
Hungary; [email protected] the Internet has the advantages of the previously used static and dynamic telepathology methods.
Accepted for publication Conclusions: Diagnostic concordance was found between OM and VM. The digital slide and the vir-
14 January 2003 tual microscope can be alternative techniques in the computerisation of the histology laboratory and in
....................... teleconsultation services after further evaluation of time and storage constraints.

A
s a result of the rapid developments in information tech-
nology, medical digital image analysis is becoming a Table 1 Diagnoses
widely accepted technique.1 Digital techniques and Organ Diagnoses No of cases
laboratories2 already exist for radiological applications. New
Gastric Healthy 5
stand alone automated microscope systems have been Chronic gastritis without intestinal metaplasia 16
developed for histological and cytological applications in the Chronic gastritis with intestinal metaplasia 12
past decade. Chronic gastritis with atrophia 3
The automated screening and rescreening of cervical Adenocarcinoma, well differentiated, intestinal 9
type
smears is now available for routine practice.3 4 In addition,
Adenocarcinoma, non-differentiated 6
several new semiautomatic microscopes have been developed Sigillocellular carcinoma 4
to aid in quality control of the cytotechnologists’ work.5–8 Anaplastic carcinoma 2
These systems notify the observer of the scanned areas and Adenopapillar carcinoma 1
images, and electronic recording of selected images is also Hyperplastic polyp 1
MALT lymphoma 2
available.
Automated histological analysis has been an important Colon Healthy 5
research tool for several years, but until recently it was not Colitis ulcerosa 14
suitable for routine applications.9 10 However, now histological Crohn’s disease 7
diagnoses can be supported by new electronic techniques, Chronic aspecific colitis 4
Hyperplastic polyp 2
such as TV image cytometry and teleconsultation based on
Adenoma tubulare with severe dysplasia 2
histological images rather than entire samples.11 Adenoma tubulovillusom with severe dysplasia 1
Adenoma papillare 1
Adenoma tubulopapillare 1
“Automated histological analysis has been an important Adenoma villosum with dysplasia 1
research tool for several years, but until recently it was Adenocarcinoma 4
not suitable for routine applications” MALT, mucosa associated lymphoid tissue.

Telepathology services were built up during the past decade

around two technology platforms. Dynamic telepathology much less hardware investment, but sampling errors can
uses remote controlled microscope systems with high occur.14 15 Using the Internet as a telecommunication pathway
throughput online image transport channels.12 13 This method for static images is a low cost widely available alternative.16 17
has the advantage of entire slide access and lacks the error
source of preselected microscopical frames, although the costs
.............................................................
of setting up and running the system are relatively high. The
use of static preselected images for teleconsultation needs Abbreviations: OM, optical microscopy; VM, virtual microscopy

www.jclinpath.com
434 Molnar, Berczi, Diczhazy, et al

J Clin Pathol: first published as 10.1136/jcp.56.6.433 on 1 June 2003. Downloaded from http://jcp.bmj.com/ on 27 May 2019 by guest. Protected by copyright.
Figure 1 The user interface of the virtual microscope program. The artificial magnification (×10) of a scanned gastric biopsy section is shown
on the screen. Note the moving arrows in the left bottom corner and the + and – labels in the middle. These can be used for discrete step
magnification and the reduction of a specified area of the slide. On the right bottom of the screen the scanned area of the slide is highlighted.

The applications of digital slides are considerable, as sources (OM), without knowledge of the results of the previous
and targets of telepathology and automated histological analysis by two independent, board certified histologists.
analysis; however, attempts to prepare electronic or digital First, the glass slides were assessed by optical microscopy.
slides have been limited (R Ferreira et al. Digital dynamic The paper copies of the clinical histories and the evaluation
telepathology—the virtual microscope. Proceedings of the report were collected.
1997 AMIA Annual Fall Symposium, 1997).18 19 Only recently Over the next few weeks the slides were digitised; this was
has the storage capacity and speed of personal computers carried out in the digital microscopy laboratory. The scanning
become sufficient to handle the extremely large amount of computer was also used as a digital slide server. The virtual
image information stored on a slide. In addition, low cost, microscopy evaluation was done on a separate local area net-
commercial motorised microscopes have recently been devel- work workstation in the pathology department with access to
oped by several manufacturers. the slide server computer using the Internet.
Working on digital slides requires virtual or digital micros-
copy. Preliminary positive results on a limited number of Data analysis
mosaicked microscopic images have been reported recently.20 At the end of the study, when all the optical and virtual
The aim of our present study was to evaluate our recently microscopy results were available, the consensus data,
reported digital slide and microscopy system21 on routine concordance, and source of diagnostic discordance were
gastrointestinal biopsy specimens. We compared the perform- determined.
ance of optical and digital microscope evaluations in local and Concordance was designated level “A”. Levels “B” and “C”
remote modes. were clinically unimportant and important discordance,
respectively, as suggested by Weinberg et al.22
MATERIALS AND METHODS Those cases with discordant results were collected together
Gastric routine biopsy specimen analysis with the optical and electronic data. The final diagnosis and
Biopsy specimens were placed in buffered formalin, routinely definition of the source of the discordant data were agreed
processed, and stained with haematoxylin and eosin. Alto- upon at a consensus session in the consultation room. Here,
gether, 103 specimens were selected from the files of the first access to the digital slides was also available through a
department of pathology. Single representative slides were computer workstation and the Internet.
evaluated from each case. Table 1 shows the distribution of the The reasons for discordance were classified by the
cases. consensus meeting as follows: inadequate image quality (class
The histological sections were evaluated in separate settings I), interpretation (class II), and insufficient clinical infor-
on a virtual microscope (VM) and an optical microscope mation (class III).

www.jclinpath.com
Digital slide and virtual microscopy based techniques 435

J Clin Pathol: first published as 10.1136/jcp.56.6.433 on 1 June 2003. Downloaded from http://jcp.bmj.com/ on 27 May 2019 by guest. Protected by copyright.
Figure 2 An intermediate magnification of a scanned gastric biopsy specimen (original magnification, ×200).

The significance of the concordance was determined using would need to be stored for the entire slide. However, a
Kendall’s concordance coefficient determined by the Statistica threshold filter was used to store only those frames with
program package (V.4.3, Statsoft, Tulsa, Oklahoma, USA). image content. In this way, only the area containing the biopsy
was stored, amounting to 302 to 1334 frames for each slide.
Slide digitisation and the virtual microscope system First, the fields of view were mosaicked using mathematical
Hardware tools used algorithms24; however, the error between the required and real
We used an Axioplan 2 MOT (Carl Zeiss, Jena, Germany) positions of the stage was found to be less than ± 0.5 µm.
microscope. The microscope functions (objectives, stage, Based on this high precision no software mosaic alignment
focus, illumination, and filters) can be controlled and changed was used.
through the RS232 interface from an application program. The
mechanical accuracy of the motorised scanning stage for X/Y Features of the virtual microscope program
and Z directions was 0.5 µm. Slide selection
We used the Grundig FAC 830 1/2 inch, one chip CCD cam- After scanning, the slides were stored in subdirectories called
era with 752 × 581 (440 000) pixels. The integration time of projects for a higher ordering. After selection an electronically
the camera can be controlled through the computer interface minimised slide image is shown on the screen (fig 1).
RS485. The applied image digitisation card was the Screen
Slide orientation map
Machine II from Fast Electronic Germany (Munich, Germany)
This map represents the whole slide, where one pixel on the
with a resolution of 640 × 560 pixels and 64 K colour depth.
screen corresponds to one field of view. The recorded segments
The programs were running on computers with at least an
are labelled with white pixels on a grey background.
Intel Pentium II 350 MHz processor, 128 MB RAM, and 2 GB
hard disk. Applicable electronic magnifications
After finding the area of interest, the user has several options
Features of the digital slide scanning program for magnification of the selected segments. The prepared
Scanning area determination was performed by upper right magnification steps of ×100 and ×200 or special “+” and “−”
and lower left corner fixing. mouse arrows can be used to change the magnification (figs
Autoscanning was started after setting the stage at the zero 1–3).
position. During the scanning process autofocusing was done
only at each third to fifth field of view. All the images were Moving and scrolling the slide
compressed in JPEG format and stored in the slide databank If the interested area is not on the screen at the required mag-
in the corresponding position. Autofocusing was done using nification, then the moving arrows (up, down, left, and right)
Brenner’s algorithm.23 At ×40 magnification, 125 856 frames can be used to move the slide in any direction.

www.jclinpath.com
436 Molnar, Berczi, Diczhazy, et al

J Clin Pathol: first published as 10.1136/jcp.56.6.433 on 1 June 2003. Downloaded from http://jcp.bmj.com/ on 27 May 2019 by guest. Protected by copyright.
Figure 3 The real world scanning magnification (original magnification, ×400) on the virtual microscope. The borders between the
mosaicked microscopic field of views can also be detected.

Labelling of interesting frames for re-evaluation, Table 2 Concordance between optical microscopy,
consultation, and reporting virtual microscopy, and the consensus diagnosis
Up to 10 coloured labels can be placed on the digital slide in
the software. This option can also be used for reconsultation Concordance (all Concordance (clinically
types) important)
by experts via the local area network or in specific cases via the
Internet. OM v VM 92% (95/103) 94% (97/103)
OM v Con 97% (100/103) 98% (101/103)
VM v Con 95% (98/103) 96% (99/103)
Internet access
Every virtual microscopy workstation can be a slide server too. Con, consensus; OM, optical microscopy; VM, virtual microscopy.
On the remote workstation the slide server’s IP address has to
be defined. After connecting to the server computer, a list of
available slides and the information about these slides is Reasons for diagnostic discordance
transported to the client workstation. After selecting a slide, Because the entire specimen was digitised there was no sam-
its electronically compressed low resolution image is trans- pling error. However, the other common sources of error in
ferred to the workstation. Every image that is transported telepathology—low image quality (one of eight), interpret-
during the evaluation will be stored on the local machine for ation (four of eight), and insufficient clinical information
safety reasons. (three of eight) were seen.
The effects of poor image quality were seen in only one case,
although this was a clinically important one. In this case, the
RESULTS VM diagnosis was ulceration ventriculi and the OM diagnosis
Concordance was ulcerated well differentiated adenocarcinoma ventriculi.
Discordance was found in eight cases (7.8%) between the VM The correct diagnosis was achieved by evaluating multiple
or OM results and the consensus diagnosis. In five cases the focus levels in OM, an option that is not currently available in
OM and in three cases the VM yielded the correct diagnosis. VM (table 3).
There was concordance in 95 (92.2%) of the 103 cases, and
clinically important concordance in 97 cases (94.1%). OM Technical and practical data regarding the application
yielded higher (100 of 103; 97%) concordance with the of VM
consensus results than did VM (98 of 103; 95.1%). The high- The hard disk volume of a microscopic field of view is between
est degree of concordance for clinically important diagnosis 60 and 100 KB after JPEG compression. The overall hard disk
was also seen for OM (101 of 103; 98%) (table 2). space required for a gastric biopsy is between 30 and 50 MB

www.jclinpath.com
Digital slide and virtual microscopy based techniques 437

Table 3 Reasons for diagnostic discordance

Review
Level of Reason for consensus

J Clin Pathol: first published as 10.1136/jcp.56.6.433 on 1 June 2003. Downloaded from http://jcp.bmj.com/ on 27 May 2019 by guest. Protected by copyright.
discordance discordance VM diagnosis OM diagnosis diagnosis

B II Chronic gastritis Acute and chronic gastritis VM

C III Gastric adenoma Gastric hyperplastic polyp VM
C I Ulceration ventriculi Ulcerated, well differentiated adenocarcinoma OM
C II Chronic atrophic gastritis Chronic gastritis with low grade dysplasia OM
B II Hyperplastic polyp Hyperplastic mucosa OM
C II Mild colitis Normal colon OM
C III Colitis ulcerosa with dysplasia Colitis ulcerosa VM
C III Adenoma papillare Hyperplastic polyp OM

Discordance: B, clinically not important; C, clinically important. Reason for discordance: I, image quality; II, interpretation; III, insufficient clinical
information.
OM, optical microscopy; VM, virtual microscopy.

Table 4 Comparison between optical and virtual microscope evaluation methods

Microscopy feature Optical evaluation Electronic evaluation

Speed 4–5 minutes 2–3 minutes

Costs Every user, high quality microscope One high quality scanning microscope, image server, PC
Working conditions Only limited number of slides can be evaluated because of Normal monitor work
fatigue
Consultation With slide transportation or with image transfer, through ISDN, Internet Remote access to the image server through local
telephone line area network or internet
Quality control With special additional hardware By recording of evaluated image segments

and the scanning time is between 20 to 40 minutes, depending We found a higher concordance with optical microscopy
on the number and area of sections on a slide. than previous static image based telepathology studies.13–15
Evaluating the specimen on the monitor is more comfort- This might be explained by the fact that this technology
able for the histologist, more reproducible, and more easily eliminates sampling error.
documented when compared with the optical method (table Time constraints and fatigue were not evaluated. We are
4). planning to assess these factors in a larger interlaboratory
Remote access via the Internet was relatively fast. The first study.
minimised image was uploaded in real time. As the user
switched over to high magnification the single fields of view “In the near future we can have microscope free virtual
were seen on the screen in seconds (figs 2,3). microscopy workstations with the functions of the optical
microscope”
DISCUSSION
In our study, we evaluated the technical feasibility and
The reasons for discordance yielded new information for
diagnostic concordance of a digital histological section evalu-
future developments. In selected cases several focus planes are
ation system and compared it with the routine OM procedure.
required for correct diagnosis. The virtual microscope cannot
Low cost computational techniques were used. The scan-
be used for the analysis of underlying cells. However, with fur-
ning speed was too slow for routine applications, but there are
ther technical and mathematical developments several focus
free resources available for increasing scanning speed by sev-
levels can be recorded, stored, and evaluated by the virtual
eral magnitudes. The digitisation card should be changed to a
higher resolution and speed (×10). The computer should microscope.
also be updated (×5). In addition, an automated microscope In his 1996 review article, O’Brien stated that the
could be considered. computerisation of the histology laboratory was desirable but
Image quality should be further enhanced using a digital that it would be far in the future.26 More, recently, Leong and
one chip, analogue three chip camera instead of the analogue, McGee27 stated that complete automated slide digitisation has
one chip camera used at present. In this case, an increase in influence at all levels of clinical practice and education. They
diagnostic accuracy would be expected. emphasised the importance of dedicated software technology.
However, these developments and results show that in the Recently, the usefulness of digital slides and virtual micros-
near future we can have microscope free virtual microscopy copy for quality assurance and teaching applications has been
workstations with the functions of the optical microscope. demonstrated.28 29
One advantage would be a more comfortable working Our results show that digital slides and virtual microscopy
environment for histologists. In addition, it would also technology can be used in selected cases for telepathology
support quality control techniques and consultation with consultation. After the definition of the time and storage
remote experts. The first application of digital slides and requirements the routine use of this technique could also be
virtual microscopy should be telepathology consultation. considered.
However, later this technology could be used in the routine
histology evaluation process, as was done in our study. .....................
Remote evaluation of the slides through the Internet has Authors’ affiliations
the advantages of the previously used static or dynamic B Molnar, A Tagscherer, S V Varga, Z Tulassay, Digital Microscopy
Laboratory, Second Department of Medicine, Semmelweis University,
telepathology methods (the entire slide is available at high Szentkirályi St. 46, 1088 Budapest, Hungary
magnification, microscope and remote assistance free L Berczi, C Diczhazy, B Szende, First Department of Pathology and
evaluation).25 Experimental Oncology, Semmelweis University

www.jclinpath.com
438 Molnar, Berczi, Diczhazy, et al

11 Dictor M. The surgical pathologist in a client/server computer network:

Take home messages work support, quality assurance, and the graphical user interface. Mod
Pathol 1997;10:259–66.
• We found diagnostic concordance between optical micro- 12 Dunn BE, Almagro UA, Choi H, et al. Dynamic-robotic telepathology:
Department of Veteran Affairs feasibility. Hum Pathol 1997;28:8–12.

J Clin Pathol: first published as 10.1136/jcp.56.6.433 on 1 June 2003. Downloaded from http://jcp.bmj.com/ on 27 May 2019 by guest. Protected by copyright.
scopy and virtual microscopy on gastrointestinal biopsy
13 Weinstein RS, Battacharayya AK, Graham AR, et al. Telepathology a
specimens ten-year progress report. Hum Pathol 1997;28:1–7.
• Virtual microscopy has the advantage of providing a more 14 Cross SS, Burton JL, Dube AK, et al. Offline telepathology diagnosis of
comfortable working environment for histologists colorectal polyps: a study of interobserver agreement and comparison
• Virtual microscopy also supports quality control techniques with glass slide diagnoses. J Clin Pathol 2002;55:452–60.
and consultation with remote experts 15 Tucker JH, Busch C, Spatz A, et al. An experimental inter-expert
• Initially, digital slides and virtual microscopy should be telepathology network using static imaging. J Clin Pathol
2002;52:761–6.
used in telepathology consultation 16 Montironi R, Thompson D, Scarpelli M, et al. Transcontinental
• After further evaluation of time and storage constraints this communication and quantitative digital histopathology via the Internet:
technology might be useful in routine histology with special reference to prostate neoplasia. J Clin Pathol
2002;55:305–13.
17 Gombas P, Szende B, Stotz G. Support by telecommunications in
diagnostic pathology. Experience with the first telepathology system in
Hungary. Orv Hetil 1996;137:2299–303.
REFERENCES 18 Burns BF. Creating low power photomicrographs using a 35 mm digital
1 Shotton DM. Robert Feulgen prize lecture 1995. Electronic light slide scanner. Am J Surg Pathol 1997;21:865–6.
microscopy: present capabilities and future prospects. Histochem Cell 19 Demichelis F, Barbareschi M, Dalla Palma P, et al. The virtual case: a
Biol 1995;104:97–137. new method to completely digitize cytological and histological slides.
2 Dooley RL, Engel C, Muller ME. Automated scanning and digitizing of Virchows Arch 2002;441:159–64.
roentgenographs for documentation and research. Clin Orthop 20 Singson RPC, Natarajan S, Greenson JK, et al. Virtual microscopy and
1997;274:113–19.
the Internet as telepathology consultation tools. A study of gastrointestinal
3 Mango LJ, Ivasauskas EZ. Computer assisted cervical cytology using the
biopsy specimens. Am J Pathol 1999;111:792–5.
PAPNET testing. In: Wied GL, Keebler CM, Rosenthal DL, et al, eds.
21 Molnar B, Papik K, Tagscherer A, et al. Three-dimensional reconstruction
Compendium on quality assurance, proficiency testing and workload
and analysis of gastric malignancies by electronic slides of consecutive
limitations on clinical cytology. Chicago, USA: Tutorials of Cytology,
sections and virtual microscopy. Proceedings of the International Society
1995:155–67.
for Optical Engineering 1999;3605:190–200.
4 Hailey DM, Lea R. Prospects for newer technologies in cervical cancer
screening programmes. J Qual Clin Pract 1995;15:139–45. 22 Weinberg DS, Allaert FA, Dusserre P, et al. Telepathology diagnosis by
5 Baker RW, Wadsworth J, Brugal G, et al. An evaluation of “rapid means of digital still images: an international validation study. Hum
review” as a method of quality control of cervical smears using the Pathol 1996;27:111–18.
AxioHOME microscope. Cytopathology 1998;8:85–95. 23 Firestone L, Cook K, Culp K, et al. Comparison of autofocus methods for
6 Grohs DH, Dadeshidze VV, Domanik RA, et al. Utility of the TracCell automated microscopy. Cytometry 1991;12:195–206.
system in mapping Papanicolaou-stained cytologic material. Acta Cytol 24 Ott SR. Acquisition of high-resolution digital images in video microscopy:
1997;41:144–52. automated image mosaicking on a desktop microcomputer. Microsc Res
7 Anderson TL, Nelson AC. Quality control and proficiency testing of Tech 1997;38:335–43.
cytological smear screening: an integrated approach using automation. 25 Weinberg DS. How is telepathology being used to improve patient care.
In: Wied GL, Keebler CM, Rosenthal DL, et al, eds. Compendium on Clin Chem 1996;42:831–5.
quality assurance, proficiency testing and workload limitations on clinical 26 O’Brien MJ, Sotnikov AV. Digital imaging in anatomic pathology. Am J
cytology. Chicago, USA: Tutorials of Cytology, 1995:83–287. Clin Pathol 1996;106:25–32.
8 Grohs DH, Gombrich PP, Domanik RA. AccuMed International, Inc. 27 Leong FJWM, McGee O’D. Automated complete slide digitization: a
Meeting the challenges in cervical cancer screening: the AcCell Series medium for simultaneous viewing by multiple pathologists. J Pathol
2000 automated slide handling and data management system. Acta 2001;195:508–14.
Cytol 1996;40:26–30. 28 Demichelis F, Della Mea V, Forti S, et al. Digital storage of glass slides
9 Ong SH, Jin XC, Sinniah R. Image analysis of tissue sections. Comput for quality assurance in histopathology and cytopathology. J Telemed
Biol Med 1996;26:269–79. Telecare 2002;8:138–42
10 Belhomme P, Elmoataz A, Herlin P, et al. Generalised region growing 29 Hediger PM, Dee F, Consoer D, et al. Integrated approach to teaching
operator with optimal scanning: application to segmentation of breast and testing in histology with real and virtual imaging. Anat Rec
cancer images. J Microsc 1996;886:41–50. 2002;269:107–19.

www.jclinpath.com