Engineering, 2013, 5, 103-107
http://dx.doi.org/10.4236/eng.2013.510B021 Published Online October 2013 (http://www.scirp.org/journal/eng)
Copyright © 2013 SciRes. ENG
Developing Customized Evaluation Software for Clinical
Trials: An Example with Obstructive Lung Diseases
Zhanqi Zhao1,2, Barbara Vogt3, Inéz Frerichs3, Ullrich Müller-Lisse2, Knut Möller1
1Institute of Technical Medicine, Furtwangen University, VS-Schwenningen, Germany
2Department of Radiology, University of Munich, Germany
3Department of Anesthesiology and Intensive Care Medicine, University Medical Center Schleswig-Holstein,
Campus Kiel, Kiel, Germany
Email: zhanqi.zhao@hs-furtwangen.de
Received December 2012
ABSTRACT
Corresponding customized software tool is usually unavailable, which increases the time and workload for evaluating
the results of a clinical trial. In the present paper, w e demonstrate the development process of a customized software for
one clinical trial on patients with obstructive lung disease. Ov er hundred patients and volunteers as contro lled were in-
cluded in the clinical trial. They were examined by spirometry and EIT in a seated position during spontaneous tidal
breathing. Subsequently, standard vital capacity maneuver and forced full expiration maneuver were performed. In or-
der to evaluate the offline data, a customized software was developed. The requirements of the software were defined
by investigators. The software was then tested on patients’ data and refined based on feedbacks of the investigators. We
finalized the customized software with analysis of various disease-specific parameters and indices. Compared to the
data process with device specific programs and other commercial software, the customized software is more flexible,
user-friendly and extendable. As conclusion, customized software simplifies the evaluation process distinctly and helps
physicians to focus on study design and result interpretation.
Keywords: Clinical Trial; Electrical Impedance Tomography; Pulmonary Function Test; Customized Software
1. Introduction
One effective way to validate a health intervention is
clinical trial. It is usually prospective, has no well estab-
lished protocol, and generates huge amount of data in
order to reach sufficient statistical power. The data often
need to be manually evaluated with commercial spread-
sheet or statistical software (e.g. Microsoft Excel, IBM
SPSS) according to our experiences. These softwares are
unspecific and not extendable for novel algorithms.
Therefore, the evaluation processes are time-consuming
and tedious for the physicians on one hand, and on the
other hand, limited to conventional analysis, which is
unwanted especially for pilot studies.
Customized evaluation software for clinical trials does
warrant under these circumstances. Such software need
to be study specific, user-friendly and extendable for
further studies. Close collaboration between physicians
and engineers is the prerequisite of the customized soft-
ware development. Recently in a series of clinical trials,
the feasibility of electrical impedance tomography (EIT)
on identifying inhomogeneous regional lung ventilation
was examined in patients with obstructive lung disease
(e.g. [1,2]). In the present paper, we report an example of
customized software for one of these clinical trials and
demonstrate the development of the software for batch-
ing EIT dat a .
2. Materials and Methods
2.1. Electrical Impedance Tomography for
Thorax
Thorax EIT is a noninvasive imaging technique used for
monitoring the regional lung ventilation and tidal volu me
distribution [3]. The rationale of EIT measurement is that
changes in regional air content and regional blood flow
alter the electrical impedance of lung tissue [4]. The in-
jected alternating current (typically chosen between 10
kHz - 100 kHz) through th e thorax result in non -uniform
distribution of electrical potentials at the chest wall sur-
face. Thorax EIT systems with 16 electrodes are widely
used, where 208 voltages are measured in each frame.
EIT images showing time-difference impedance distribu-
tion within measuring layer are reconstructed based on
the voltage measurement. Applications of EIT have been
proposed to assess the ventilation distribution or guiding
respiratory therapies in subjects without lung disease [5,
Z. Q. ZHAO ET AL.
Copyright © 2013 SciRes. ENG
104
6] and in patients with mild, moderate, and severe lung
diseases [7-9]. However, evaluation of EIT measurement
is not easy for doctors since 1) there is no established
guideline for measurement; 2) commercially available
software is device specific and rarely extensible for var-
ious clinical trials; 3) open source software is compre-
hensive but not disease specific.
We recently developed a customized EIT evaluation
software for a clinical study in university hospital
Schleswig-Holstein. The examination protocol of the
clinical trial is briefly described, followed by the devel-
opment process of the software.
2.2. Examination Protocol
At the time point of writing the manuscript, over one
hundred patients with chronic obstructive pulmonary
disease (COPD), cystic fibrosis (CF), or asthma were
included in the clinical trial. All subjects were examined
by spirometry (Jaeger pneumotachograph, CareFusion,
Höchberg, Germany) and EIT (Goe-MF II EIT system,
CareFusion, Höchberg, Germany) in a seated position
during spontaneous tidal breathing. They were asked to
perform full inspiration from functional r esidual capacity
to total lung capacity, followed by standard forced full
expiration maneuver and at the end back to spontaneous
tidal breathing (F ig ure 1) [2]. For the EIT measurement,
16 ECG electrodes were placed on the chest circumfe-
rence in the 5 - 6th intercostal space (parasternal line)
and one reference electrode on the abdomen in each sub-
ject. Data acquisition lasted about 60 - 100 s.
2.3. Development of the Evaluation Software
Software requirements and design: The engineers from
Furtwangen University and the clinicians from Universi-
ty Medical Center Schleswig-Holstein have held a series
of meetings to discuss the settings of the clinical trial; to
elicit, analyze, specify and validate the requirements of
the evaluation software. Illustrations of various disease-
specific parameters and indices were graphically summa-
rized by the clinicians (e.g. Figure 2(A)). Evaluation
requirements were described in functional flow block
diagram (Figure 2(B)).
Software construction: Considering that a very useful
EIT open source tool EIDORS [10] is programmed with
MATLAB (The MathWorks, Natick, MA, USA), the
evaluation software is written with MATLAB (Ver. 7.2)
and built with MATLAB Deploy tool and C++ Compiler.
Software testing and refinem ent : The behavior and
performance of the evaluation software was dynamically
verified on a set of test cases before delivered to the phy-
sicians. Feedback was given by the physicians after test-
ing the software on patient data. Unexpected behaviors of
the software due to misunderstanding or bugs were cor-
Figure 1. Spirogram of lung volume changes during the
examination. Four phases are recognized: spontaneous tidal
breathing, vital capacity maneuver, forced expiration ma-
neuver and again spontaneous tidal breathing. VT: tidal
volume; ΔEELV: change of end-expiratory lung volume.
rected. This testing-improving process was repeated sev-
eral times to improve the software.
3. Results
The software consists of two parts: The physicians use the
first part to process the raw measurement data (Figure
3(A)). Five different EIT raw and image data formats can
be recognized. EIT images are reconstructed with either
back projection or GREIT method and can be extended
easily. Fifty frequently used clinical parameters such as
forced expiratory volume in 1 second (FEV1) and peak
expiratory flow (PEF) and their quotients can be quickly
calcula te d by c licking the c or r e s ponding butt ons . EIT data
and results are graphically displayed and as different for-
mats exported or printed (e.g. JPEG, TXT, XLS, PDF;
Figure 3(B)). Simple filters such as moving average and
butterworth are implemented to suppress signals other
than respiration.
In the second part of the software, various disease-spe-
cific parameters and indices can b e displayed as function-
al images or histogram in different scales and combina-
tions (Figure 4). Comparison of the parameters between
measurements, among different maneuvers or among dif-
ferent patients can be easily conducted via choosing the
corresponding options. Several validity checks (e.g. the
study group, the input ranges) are performed to prevent
mistakes of data evaluation.
4. Discussion and Conclusions
In the present study, we demonstrated the development
process of a customized evaluation software for a clinical
trial of pulmonary function test with EIT. The custo-
mized software has the following advantages compared
to the data process with device specific program and other
Z. Q. ZHAO ET AL.
Copyright © 2013 SciRes. ENG
105
Figure 2. Requirements of the evaluation software. (A) illustration of parameters in pulmonary function test. Corresponding
parameters in EIT data should be analyzed; (B) block diagram of the work flow.
commercial spreadsheet software: it is able to read dif-
ferent data formats and generate EIT images with differ-
ent reconstruction methods; it analyzes and displays var-
ious disease-specific parameters and indices in functional
images and histograms; it is user-friendly and extenda-
ble.
Up to now, most of the EIT analyses were made by
comparison between simple geometric region of interests
[11,12]. Lack of analysis tools limited the process of the
data and restricted new finding in the results. With the
disease-specific EIT evaluation software, the clinicians
are able to assess EIT images pixel-by-pixel and imple-
ment new analysis ideas. Besides, the presented software
simplifies the evaluation process distinctly and helps
physicians to focus on study design and result interpreta-
tion. The software can be easily modified and extended
to apply on patients with other lung diseases.
The experiences collected in developing the present
Z. Q. ZHAO ET AL.
Copyright © 2013 SciRes. ENG
106
Figure 3. The first part of evaluation software for processing raw measurement data. (A) interface. (B) as document for both
investigator and subject, the measurement results consisting of important findings are ready for print.
Figure 4. The second part of the evaluation software for processing various disease-specific parameters and indices. The se-
lected parameters are graphically displayed as histograms and functional images.
software are transferable and applicable for other clinical
trials. In the early stage of the development, it is very
important for the engineers to understand the problem
and background of the clinical trials. The physicians, on
the other hand, have to define the tasks and specify the
requirements in a way that the engineers can understand.
Communication and discussion between the physicians
and engineers are crucial in the stage of software re-
quirement and design. Once the developer version of the
software is ready, the engineers should test the software
function together with the physicians and finalize the
software accordingly.
A limitation of the paper is that we demonstrated the
development process only on an observational study. In
an interventional study, the need of evaluation software
may be different. In order to develop customized soft-
ware more efficiently, more experiences are needed.
5. Acknowledgements
This work was partially supported by Bundesministerium
für Bildung und Forschung (BMBF, FKZ 01|B10002D,
WiMVENT).
REFERENCES
[1] Z. Zhao, et al., “Regional Ventilation in Cystic Fibrosis
Measured by Electrical Impedance Tomography,” Jour-
nal of Cystic Fibrosis, Vol. 11, 2012, pp. 412-418.
Z. Q. ZHAO ET AL.
Copyright © 2013 SciRes. ENG
107
http://dx.doi.org/10.1016/j.jcf.2012.03.011
[2] B. Vogt, et al., Spatial and Temporal Heterogeneity of
Regional Lung Ventilation Determined by Electrical Im-
pedance Tomography during Pulmonary Function Test-
ing,” Journal of Applied Physiology, Vol. 113, 2012, pp.
1154-1161.
http://dx.doi.org/10.1152/japplphysiol.01630.2011
[3] Z. Zhao, et al., Evaluation of an Electrical Impedance
Tomography-Based Global Inhomogeneity Index for Pul-
monary Ventilation Distribution,Intensive Care Medi-
cine, Vol. 35, 2009, pp. 1900-1906.
http://dx.doi.org/10.1007/s00134-009-1589-y
[4] B. H. Brown, et al., Cardiac and Respiratory Related
Electrical Impedance Changes in the Human Thorax,
IEEE Transactions on Biomedical Engineering, Vol. 41,
1994, pp. 729-734.
http://dx.doi.org/10.1109/10.310088
[5] I. Frerichs, et al., “Gravity Effects on Regional Lung
Ventilation Determined by Functional EIT during Para-
bolic Flights,” Journal of Applied Physiology, Vol. 91,
2001, pp. 39-50.
[6] Z. Zhao, et al., PEEP Titration Guided by Ventilation
Homogene ity: A Feasibil ity Study Using Elec trical Impe-
dance Tomography,” Critical Care, Vol. 14, 2010, p. R8.
http://dx.doi.org/10.1186/cc8860
[7] E. L. Costa, et al., Real-Time Detection of Pneumotho-
rax Using Electrical Impedance Tomography,” Critical
Care Medicine, Vol. 36, 2008, pp. 1230-1238.
http://dx.doi.org/10.1097/CCM.0b013e31816a0380
[8] B. M. Eyuboglu, et al., Application of Electrical Imped -
ance Tomography in Diagnosis of Emphysema: A Clini-
cal Study,Physiological Measurement, Vol. 16, 1995,
pp. A191-A211.
http://dx.doi.org/10.1088/0967-3334/16/3A/018
[9] K. Lowhagen, et al., “Regional Intratidal Gas Distribution
in Acute Lung Injury and Acute Respiratory Distress
Syndrome-Assessed by Electric Impedance Tomography,”
Minerva Anestesiologica, Vol. 76, 2010, pp. 1024-1035.
[10] A. Adler and W. R. Lionheart, Uses and Abuses of EI-
DORS: An Extensible Software Base for EIT,Physio-
logical Measurement, Vol. 27, 2006, pp. S25-42.
http://dx.doi.org/10.1088/0967-3334/27/5/S03
[11] D. Steinmann, et al., Electrical Impedance Tomography
to Confirm Correct Placement of Double-Lumen Tube: A
Feasibility Study,British Journal of Anaesthesia, Vol.
101, 2008, pp. 411-418.
http://dx.doi.org/10.1093/bja/aen166
[12] I. G. Bikker, et al., “Lung Volume Calculated from Elec-
trical Impedance Tomography in ICU Patients at Differ-
ent PEEP Levels,Intensive Care Medicine, Vol. 35,
2009, pp. 1362-1367.
http://dx.doi.org/10.1007/s00134-009-1512-6