Clinical map document based on XML (cMDX): document architecture with mapping feature for reporting and analysing prostate cancer in radical prostatectomy specimens
© Eminaga et al; licensee BioMed Central Ltd. 2010
Received: 26 May 2010
Accepted: 15 November 2010
Published: 15 November 2010
The pathology report of radical prostatectomy specimens plays an important role in clinical decisions and the prognostic evaluation in Prostate Cancer (PCa). The anatomical schema is a helpful tool to document PCa extension for clinical and research purposes. To achieve electronic documentation and analysis, an appropriate documentation model for anatomical schemas is needed. For this purpose we developed cMDX.
The document architecture of cMDX was designed according to Open Packaging Conventions by separating the whole data into template data and patient data. Analogue custom XML elements were considered to harmonize the graphical representation (e.g. tumour extension) with the textual data (e.g. histological patterns). The graphical documentation was based on the four-layer visualization model that forms the interaction between different custom XML elements. Sensible personal data were encrypted with a 256-bit cryptographic algorithm to avoid misuse. In order to assess the clinical value, we retrospectively analysed the tumour extension in 255 patients after radical prostatectomy.
The pathology report with cMDX can represent pathological findings of the prostate in schematic styles. Such reports can be integrated into the hospital information system. "cMDX" documents can be converted into different data formats like text, graphics and PDF. Supplementary tools like cMDX Editor and an analyser tool were implemented. The graphical analysis of 255 prostatectomy specimens showed that PCa were mostly localized in the peripheral zone (Mean: 73% ± 25). 54% of PCa showed a multifocal growth pattern.
cMDX can be used for routine histopathological reporting of radical prostatectomy specimens and provide data for scientific analysis.
To our knowledge, there is no established electronic standard for graphical documentation of PCa that meets clinical and research requirements. These requirements include a flexible documentation of PCa extension with an anatomical schema that can be used for clinical and research purposes. To provide an analyzable data acquisition model for anatomical schemas in electronic form, we propose a document architecture called "clinical Map Document based on XML" (cMDX). The development of this document architecture depends on clear definitions of domain terminologies, functional and data hierarchies, as well as assessment rules. It is intended to improve information consistency and data integrity of routine data in hospital information systems and clinical studies.
This article describes a data model based on schematic diagrams for documentation and analysis of prostatectomy specimen reports.
List of presentation attributes.
If true, then fill the shape otherwise not.
If true, show only the intersected drawing surface with mask.
Specifies the type of the brush (Hatchbrush, Solidbrush)
Set the colour opacity of filled area.
Set the fill colour (Alpha;Red;Blue;Green)
180; 255; 0; 0
Specifies the alignment of lines in relation to the theoretical, zero-width line
Setting the opacity of lines (0-255)
Setting the colour of lines (A, R, B, G)
150; 0; 0; 0
Specifies the style of dashed lines
Specifies the available cap styles of the beginning of a line.
Specifies the available cap styles of the end of a line.
Specifies how to join consecutive line segments in a figure
Specifies the type of fill uses to fill lines.
Setting the width of a line
Setting the coordination of a polygon (X1; Y1 X2; Y2...)
Setting the boundary of a shape (X, Y, Width, Height)
Setting the coordination of border
Setting the boundary of border
Setting the colour of border lines
0;; 0; 0
Set if the border should be shown.
Stored binary data of the image
List of attributes for textual information: the "Befund" element describes histological patterns of the PCa.
The root element:
Identification number of PCa area
Show the Gleason score in the pathology report
Primary Gleason pattern
Secondary Gleason pattern
The Gleason Score
The size of the PCa area (Optional)
The thickness of the slice (Optional)
List of attributes for textual information: the "Form" element.
The root element:
pT, pN, pM
Date of the report
Atypical changes in cell nucleus
Grade of cell differentiation
Classification of adenocarcinoma
Primary Gleason pattern
Secondary Gleason pattern
Total Gleason score
Lymph vessel invasion
Venous vessel invasion
Status of resection margin
Number of lymph nodes with metastases
Total number of extirpated lymph nodes
Prostate diameters (Optional)
Defining the barcode type (e.g. CODE128A)
Analysis of the pathology report
The report of radical prostatectomy specimens (Figure 1) contains text and graphic data. Personal and pathologic information are textual. The anatomical scheme of the prostate gland is a template for the graphical documentation of histopathological findings. As previously published by Bettendorf et al. , the anatomical schema consists of both seminal vesicles and the prostate sectioned into eight defined slices. For each slice, a "slice factor" is attributed to estimate the tumour volume in relation to the total prostate volume. Symbols and icons are added to facilitate identification of the pathological findings. The report template was primarily designed for the documentation of PCa and High Grade Prostatic Intraepithelial Neoplasia (HGPIN), a presumable precursor for PCa [8, 9]. PCa was graded according to Gleason  and Helpap  and staged using the TNM staging system (2002) .
The TNM-classification system  is a well-known pathological documentation system that codes tumour spread (T), lymph node invasion (N) and metastases (M). Parameters like residual (R), venous invasion (V) and lymph vessel invasion (L) provided additional clinical information about PCa. The Gleason score system is a standard applied to assess the histological pattern and spread of PCa ; Helpap is a method for the histological evaluation of cell differentiation and the existence of atypical prostatic cells .
Document architecture of "cMDX"
cMDX document architecture was designed to meet the Open Packaging Convention (OPC)  (Figure 2 and additional file 1). Data were divided into template data and patient data (Figure 2 and 3). Each XML element in cMDX was declared by a corresponding class library. The element declaration was performed with the XML namespace "xmlns".
Template data (TD)
Explanatory attributes for pathological changes like Description, Name of a disease written in short form, Image-Filename for the symbolic representation of the disease, and Text for the full name of the disease.
In addition, the optional Properties Element enables the interaction of the drawing tools with the application.
Mask.xml stores configuration data to define the appearance of the graphical template including a background Image, which is stored in BGImage, and vector shapes with clip function. These shapes represent the drawing surfaces onscreen. Three shapes were defined: (1) Polygon (2) Rectangle (3) Ellipse. Every shape element, whose tag name is the applied shape name, was stored inside the root element Mask. For instance, the element Ellipse represents a slice of the prostate. It has identification (e.g. Id, Name) and representation attributes (Table 1), which, for example, provide information about the Percentage volume of the slice in relation to the whole prostate volume (slice factor), the PixelVolume of the slice, and the SizeOfSlice for storing the real dimension of the prostate slice.
The component "BGImage" includes an Image representing the prostate and the seminal vesicles in schematic style and can be changed by the user. Various image formats (TIFF, BMP, WMF, PNG, JPEG, GIF and SVG) are hereby supported.
Patient Data (PD)
Data acquisition from the prostatectomy specimen consists of two types of information: (1) morphometrical information about histological changes (e.g. PCa, HGPIN, positive surgical margin) with additional textual descriptions (Gleason Score, length of positive surgical margin), and (2) clinical and personal data in text form. Accordingly, two xml containers were designed: Form.xml and Map.xml. Map.xml captures morphometrical information about histological findings to reconstruct them in schematic styles. Diverse graphic objects were defined and stored in the root xml element Map (Figures 3 and 5). The representation of the spread pattern of PCa and HGPIN was based on vector graphics. The real dimension of a cancer area (Length and Width) as well as the slice thickness (SliceThick) can be optionally given for special scientific queries. In addition to the representation attributes, attributes like Relation and RelationSlide provide the topographical location of PCa foci in relation to each other. An XML Element called Befund was integrated into the CDATA-Section of a graphic object showing a PCa focus. Befund captures information about histological patterns of the corresponding PCa focus (Table 2).
Form.xml contains the element Formular with attributes containing personal data; pathological findings were categorised as textual information not directly related to the morphometrical data like prostate length, width and volume (ProstatLength, ProstatWidth, Prostatvolumen) (Table 3). Sensible personal data was encrypted with Rijndael 256-bit Cryptography  to avoid misuse by unauthorized users (Figure 5).
A Four-Layer model for visualization of graphical information was designed (Figure 6): (1) The Input Layer registers the cursor movement onscreen and the drawing tools selected by the user for data acquisition. (2) The Mask Layer defines the drawing surface. (3) The Draw Layer represents the drawing elements resulting from the Layers 1 and 2. (4) The Background Layer contains schematic diagrams with graphical descriptions as images.
Clinical evaluation of cMDX for estimating tumour volume and multifocal PCa
Two-hundred and fifty-five patient records were chosen randomly for a retrospective study, in which their reports of radical prostatectomy specimens were digitalized and converted into cMDX reports. We analysed PCa foci within the prostate boundary. Therefore, a C#-based tool was implemented to analyse cMDX documents. Results including mapping, tumour volume, multifocality and incidence of PCa in each slice were exported as CSV files and analysed with PASW© Statistics version 18 (SPSS Inc., Chicago, United States). At the same time, an experienced pathologist estimated tumour volume and multifocality of PCa by eyeball judgment alone. These results were then compared with the results of the analysis tool.
Cancer volume estimate
In the pathology institute the prostate volume (Prostatvolumen) was calculated after formalin fixation by weighing the prostate specimen without the seminal vesicles. For the purpose of our study, the prostate weight in grams is considered roughly equivalent to its volume in cubic centimeters (cm3); the tumour/entire gland ratio is then used to calculate the volume of the tumour in cm3. In addition, the diameters of prostate specimens were recorded (ProstatLength, ProstatWidth). A correction factor for tissue shrinkage after formalin fixation was not applied. The computational tumour volume estimate was performed on the basis of the Bettendorf's scheme : The tumour area in each slice in the diagram is estimated by counting the pixels which contain PCa. The cancer area is divided by the slice area (pixel) and then multiplied with the slice factor to calculate the relative cancer volume. Finally, the total relative cancer volume is multiplied with the prostate volume to estimate the cancer volume in cm3.
List of drawing tools.
High grade Prostatic Intraepithelial Neoplasia
Clip (TypeOfClip: Polygon)
Clip (TypeOfClip: Polygon)
positive surgical margin
Clinical Evaluation of cMDX
The cMDX documentation system was successfully connected to the local electronic Hospital Information System (HIS). The cMDX Editor can generate reports in Portable Document File (PDF) format to be imported into the corresponding electronic patient record. The complete electronic documentation took on average about eleven minutes (mean: 11 ± 2 min STD). The electronic documents were preferred by pathologists and urologists because of the clear and unambiguous presentation of the pathological findings.
The spread pattern of PCa in PZ.
Section Level of Prostate
PC Localized in PZ
73% +/- 25
Multifocality of PCa & detection rate of multifocal PCa.
cMDX is a data acquisition model for graphical and textual information. Sensible data were encrypted with 256-bit cryptographic algorithms to enable data analysis compliant with data privacy regulations. The visualization model based on the Four-Layer Model combined a graphical overview of tumour growth in a schematic style with the related textual information. In this manner, the electronic documentation of histopathological information usually given by reports in paper form is realised.
XML enables to extract clinical data and graphical schemes, and to transmit them into different formats like binary graphics (PNG, JPEG) vector graphics (SVG), Portable Document Files (PDF), text files (CVS), and Client Report Definition (rdlc) . Therefore, cMDX can be converted into a data format supported by the HIS. The file architecture is compatible with open package conventions and therefore standard documents like HL7 CDA can be integrated into cMDX .
The software prototype of the cMDX Editor can generate and modify pathology reports according to the clinical requirements of pathologists and urologists. The integration of reports of radical prostatectomy specimens into the HIS improves clinical utilities and acceptance by the clinicians.
The application provided an estimate of the tumour volume, which may be an important prognostic indicator for predicting prostate cancer recurrence following surgery . In the literature several methods to determine tumour volume of PCa can be found. These include visual estimation, computer planimetry with computer-assisted image analysis. Computer planimetry of the full set of slides of a serially sectioned prostate is believed to be the best technique for an accurate measurement of tumour volumes; however, this technique is expensive and time consuming and is therefore not applied in routine clinical practice . The estimation of tumour volume by eyeball judgment is simple and less time consuming, and cheap but shows less precision of measurement , it is performed directly under the microscope or after the transformation of tumour areas into the reconstruction map of the prostate in a schematic style . Our results show that the estimation by eyeball judgment alone resulted in lower tumour volumes as compared with cMDX estimation. The computerised estimation of tumour volume with the cMDX Editor seems to be more accurate than the estimation by eyeball judgment. We anticipate an increase in accuracy of eyeball judgement with a grid method getting close to the accuracy of the computational volume estimation. Nevertheless further investigations are needed here. However, these estimation methods are not accurate in comparison to the computer planimetry because they exclude the real dimensions of the prostate and the PCa foci.
The anatomical representation of the prostate in schematic style facilitates collecting and analysing the spread pattern of PCa in the prostate. For instance, such information can be helpful to assess the biopsy strategy in order to increase the detection rate of PCa [19–21]. Our approach enables to analyse cMDX reports and export the results into programs like SPSS© or Excel©. Our results confirm the consensus that the major tumour mass of PCa is located in the peripheral zone (PZ) . According to Chen et al., 74% of PCa foci were localized in the PZ and toward the base, diverged laterally . We found a significant correlation between tumour volume and pathological stage of the specimens, thus confirming the findings of Nelson et al. . The current analysis tool can detect multifocal PCa with a sensitivity of 90.3%. The multifocality rate of PCa was 52.5%, which is consistent with prior series that reported multifocality rates of 50 to 87% . The accuracy to differentiate multifocal from unifocal PCa depends on the applied algorithm. The transformation of the tumour area into a rectangular shape increases the size of the tumour area and increases the probability of one focus to overlap with another adjacent focus in the vicinity, this way reducing the detection rate of multifocal PCa.
A limitation of our documentation model is the documentation time necessary, cMDX failed to shorten the documentation time. cMDX is still developing and we will focus on this drawback in future versions.
An electronic documentation standard for pathologic findings of prostate specimens has not yet been introduced. We conducted a Pubmed search using the key words "electronic documentation prostate", "electronic report prostate" and "electronic documentation prostatectomy." Only one software program similar to the cMDX Editor could be found: PixelProstate (freeware), developed by Nickels , mainly focuses on the measurement of tumour volume. PixelProstate has an internal database to store clinical data and provides a simplified 3 D illustration of the prostate and PCa foci. To document the patterns of the tumour spread, the PCa foci are drawn using multiple circles. The documentation of histological patterns of each PCa focus is not supported. In contrast, the cMDX Editor enables to convert a pathologic report into a portable file format and to draw PCa foci with the freehand drawing tool. In addition, capsular invasion, extracapsular extension, positive surgical margins, and histological patterns of a PCa focus can be documented, but a 3 D illustration model similar to that of PixelProstate is not available. cMDX is a free open source document architecture. The applications based on cMDX (cMDX Editor and Analysis tool) are still prototypes (Beta-Version). We plan to develop two versions of these applications (commercial and non-commercial) and make them available as soon as possible.
cMDX offers an extensible document architecture by adding a new drawing tool and by providing additional class libraries. If needed, additional pathological parameters or characteristics can be added. For instance, recording the Gleason score of a PCa focus is feasible by adding the class library "eMocS.Definition.Befund" from the corresponding XML part of cMDX. Furthermore, this feature could extend the spectrum of the clinical applications of cMDX to various medical applications, e.g. for protocols for cardiac catheterisation or extension estimates of burn injuries in emergency surgery.
cMDX can be applied to store clinical data in schematic styles for reporting and analysing pathological parameters in radical prostatectomy specimens. It facilitates to evaluate routine data for scientific purposes.
Special thanks are due to Christiane Ptok for the continuous support of the project organisation.
- Jemal A, Siegel R, Ward E, Hao Y, Xu J, Murray T, Thun MJ: Cancer statistics, 2008. CA Cancer J Clin. 2008, 58 (2): 71-96. 10.3322/CA.2007.0010.View ArticlePubMedGoogle Scholar
- Montironi R, van der Kwast T, Boccon-Gibod L, Bono AV, Boccon-Gibod L: Handling and pathology reporting of radical prostatectomy specimens. Eur Urol. 2003, 44 (6): 626-636. 10.1016/S0302-2838(03)00381-6.View ArticlePubMedGoogle Scholar
- Epstein JI, Amin M, Boccon-Gibod L, Egevad L, Humphrey PA, Mikuz G, Newling D, Nilsson S, Sakr W, Srigley JR, Wheeler TM, Montironi R: Prognostic factors and reporting of prostate carcinoma in radical prostatectomy and pelvic lymphadenectomy specimens. Scand J Urol Nephrol Suppl. 2005, 216 (216): 34-63. 10.1080/03008880510030932.View ArticlePubMedGoogle Scholar
- McCarthy JF, Catalona WJ: Nerve-sparing radical retropubicprostatectomy. Textbook of operative urology. Edited by: Marshall FF. 1996, Philadelphia: W.B. Saunders Company, 537-544. 1Google Scholar
- Srigley JR, Amin MB, Bostwick DG, Grignon DJ, Hammond ME: Updated protocol for the examination of specimens from patients with carcinomas of the prostate gland: a basis for checklists. Cancer Committee. Arch Pathol Lab Med. 2000, 124 (7): 1034-1039.PubMedGoogle Scholar
- Bettendorf O, Oberpenning F, Kopke T, Heinecke A, Hertle L, Boecker W, Semjonow A: Implementation of a map in radical prostatectomy specimen allows visual estimation of tumor volume. Eur J Surg Oncol. 2007, 33 (3): 352-357. 10.1016/j.ejso.2006.11.004.View ArticlePubMedGoogle Scholar
- Epstein JI, Srigley J, Grignon D, Humphrey P, Association of Directors of Anatomic and Surgical Pathology: Recommendations for the reporting of prostate carcinoma. Hum Pathol. 2007, 38 (9): 1305-1309. 10.1016/j.humpath.2007.05.015.View ArticlePubMedGoogle Scholar
- Bostwick DG, Amin MB, Dundore P, Marsh W, Schultz DS: Architectural patterns of high-grade prostatic intraepithelial neoplasia. Hum Pathol. 1993, 24 (3): 298-310. 10.1016/0046-8177(93)90041-E.View ArticlePubMedGoogle Scholar
- Zynger DL, Yang X: High-grade Prostatic Intraepithelial Neoplasia of the Prostate: The Precursor Lesion of Prostate Cancer. Int J Clin Exp Pathol. 2009, 2 (4): 327-338.PubMedGoogle Scholar
- Gleason DF, Mellinger GT: Prediction of prognosis for prostatic adenocarcinoma by combined histological grading and clinical staging. J Urol. 1974, 111 (1): 58-64.PubMedGoogle Scholar
- Helpap B: Morphologic and cell kinetic studies of prostate cancers. Contribution to grading. Urol Int. 1985, 40 (1): 36-42. 10.1159/000281030.View ArticlePubMedGoogle Scholar
- Sobin LH, Wittekind C: TNM-classification of malignant tumours. 2002, Wiley-Liss, 172-175.Google Scholar
- Office OpenXML Overview. [http://www.ecma-international.org/news/TC45_current_work/OpenXML%20White%20Paper.pdf]
- AES Proposal: The Rijndael Version 2. [http://csrc.nist.gov/archive/aes/rijndael/Rijndael-ammended.pdf]
- Wang H, Azuaje F, Jung B, Black N: A markup language for electrocardiogram data acquisition and analysis (ecgML). BMC Med Inform Decis Mak. 2003, 3 (1): 4-10.1186/1472-6947-3-4.View ArticlePubMedPubMed CentralGoogle Scholar
- Using Office Open XML Formats to Support Electronic Health Records Portability and Health Industry Standards. [http://msdn.microsoft.com/en-us/library/bb879915.aspx#Office2007XMLinHealth_Overview]
- Renshaw A: Visual estimate of the percentage of carcinoma is an independent predictor of prostate carcinoma recurrence after radical prostatectomy. Cancer. 2002, 94 (8): 2310-1. 10.1002/cncr.10466. author reply 2311-2View ArticlePubMedGoogle Scholar
- Werahera PN, Miller GJ, Torkko K, Crawford ED, Stewart JS, Deantoni EP, Miller HL, Lucia MS: Biomorphometric analysis of human prostatic carcinoma by using three-dimensional computer models. Hum Pathol. 2004, 35 (7): 798-807. 10.1016/j.humpath.2004.02.015.View ArticlePubMedGoogle Scholar
- Chen ME, Troncoso P, Johnston DA, Tang K, Babaian RJ: Optimization of prostate biopsy strategy using computer based analysis. J Urol. 1997, 158 (6): 2168-2175. 10.1016/S0022-5347(01)68188-6.View ArticlePubMedGoogle Scholar
- Bolenz C, Gierth M, Grobholz R, Kopke T, Semjonow A, Weiss C, Alken P, Michel MS, Trojan L: Clinical staging error in prostate cancer: localization and relevance of undetected tumour areas. BJU Int. 2009, 103 (9): 1184-1189. 10.1111/j.1464-410X.2008.08243.x.View ArticlePubMedGoogle Scholar
- Presti JC: Prostate biopsy: how many cores are enough?. Urol Oncol. 2003, 21 (2): 135-140.View ArticlePubMedGoogle Scholar
- McNeal JE: Origin and development of carcinoma in the prostate. Cancer. 1969, 23 (1): 24-34. 10.1002/1097-0142(196901)23:1<24::AID-CNCR2820230103>3.0.CO;2-1.View ArticlePubMedGoogle Scholar
- Chen ME, Johnston DA, Tang K, Babaian RJ, Troncoso P: Detailed mapping of prostate carcinoma foci: biopsy strategy implications. Cancer. 2000, 89 (8): 1800-1809. 10.1002/1097-0142(20001015)89:8<1800::AID-CNCR21>3.0.CO;2-D.View ArticlePubMedGoogle Scholar
- Nelson BA, Shappell SB, Chang SS, Wells N, Farnham SB, Smith JA, Cookson MS: Tumour volume is an independent predictor of prostate-specific antigen recurrence in patients undergoing radical prostatectomy for clinically localized prostate cancer. BJU Int. 2006, 97 (6): 1169-1172. 10.1111/j.1464-410X.2006.06148.x.View ArticlePubMedGoogle Scholar
- Meiers I, Waters DJ, Bostwick DG: Preoperative prediction of multifocal prostate cancer and application of focal therapy: review 2007. Urology. 2007, 70 (6 Suppl): 3-8. 10.1016/j.urology.2007.06.1129.View ArticlePubMedGoogle Scholar
- Nickels J: PixelProstate: a simple software program for the measurement of prostate cancer volume. Histopathology. 2007, 50 (4): 519-521. 10.1111/j.1365-2559.2007.02602.x.View ArticlePubMedGoogle Scholar
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1472-6947/10/71/prepub