Informatics in Medicine Unlocked 6 (2017) 12–27

Contents lists available at ScienceDirect

Informatics in Medicine Unlocked

journal homepage: www.elsevier.com/locate/imu

Trends in quantitative methods used for atrial fibrillation and ventricular MARK
tachycardia analyses
⁎
Edward J. Ciaccio , Angelo B. Biviano, Vivek Iyer, Hasan Garan
Department of Medicine – Division of Cardiology, Columbia University, New York, USA

A R T I C L E I N F O A BS T RAC T

Keywords: Background: Improved quantitative and computational research efforts would be useful for better and more
Atrial fibrillation accurate analysis of heart arrhythmias, and to target catheter ablation sites. To pinpoint useful and leading-edge
MEDLINE quantitative methods, research trends of articles published in peer-reviewed journals were identified.
Publishing Methods: The MEDLINE search tool and an in-house developed software program were used to detect
Quantitative
quantitative trends in arrhythmia research. The main keywords used were ‘atrial fibrillation’ and ‘ventricular
Research trends
Ventricular tachycardia
tachycardia’, which were searched in combination with commonly associated quantitative keywords for signal
and imaging data. The search period used was 1960–2013. The linear regression trend over the search period
was calculated, and the slope and regression coefficient was tabulated along with the onset year of the trend.
Results: In 1960, ‘atrial fibrillation’ and ‘ventricular tachycardia’ appeared in the title or abstract of less than 20
peer-reviewed articles each. A sharp increase in ventricular tachycardia publications occurred from 1975 to
1992 to a peak of over 600 publications; since 1992 the number of ventricular tachycardia studies has leveled.
However, the number of atrial fibrillation papers has increased sharply since 1978, surpassing ventricular
tachycardia studies in 1993, to over 3500 studies in 2013. From 1960 to 2013, the fraction of ventricular
tachycardia papers associated with any particular quantitative keyword, versus the total number of ventricular
tachycardia publications, was often greater than the fraction of atrial fibrillation papers associated with the same
quantitative keyword, versus the total number of atrial fibrillation publications. Studies published in the
bioengineering and bioinformatics literature comprise approximately 10% of all quantitative biomedical studies
published on atrial fibrillation and ventricular tachycardia.
Conclusions: The MEDLINE search tool identified publication trends in quantitative arrhythmia research.
Although ventricular tachycardia is a common ailment, as a quantitative research topic it appears only 20% as
often as atrial fibrillation. Much of the quantitative ventricular tachycardia literature peaked in the 1990's, while
there is a continued sharp increase in atrial fibrillation research studies. This may have resulted from the
publishing of seminal studies spurring interest in quantitative analysis of atrial fibrillation, versus the
publishing of curative methods for treatment ventricular tachycardia, particularly during the period of
approximately 1995–2000.

1. Introduction
Treatment of the most common cardiac arrhythmias, specifically
atrial fibrillation (AF) and ventricular tachycardia (VT), is a subject of
intense study, with treatment guidelines appearing in the medical
literature being frequently updated [1,2]. For classification purposes,
AF is generally categorized as either paroxysmal, longstanding persis-
tent, and permanent [3], while VT is generally referred to and
discerned as monomorphic or polymorphic [4]. Quantitative character-
istics have been useful to distinguish the type of AF, and the transition
from one type to another is likely indicative of a remodeling process in
the AF substrate over time [5-7]. Catheter ablation can improve
outcome [8]. Likewise for VT, quantitative methods have been shown
useful to discern monomorphic and polymorphic types [9], which is
important to develop treatment options. Quantitative analysis of AF
can be important to localize regions for substrate modification with the
possibility of complete elimination of arrhythmogenic regions [7].
Quantitative methods for analysis of AF signals include the measure-
ment of time series [8], spectral estimation [10], nonlinear methods
[11], entropy [12], wavelet transformation [13], digital filtering [14],
and eigenanalysis [15], as well as the use of multielectrode acquisition
devices such as basket catheters [16]. Similarly, a variety of quantita-

⁎
Correspondence to: Columbia University, 630 West 168th Street, P & S 7-446, New York 10032, USA.
E-mail address: ciaccio@columbia.edu (E.J. Ciaccio).

http://dx.doi.org/10.1016/j.imu.2016.12.003
Received 3 November 2015; Received in revised form 3 November 2016; Accepted 15 December 2016
Available online 16 December 2016
2352-9148/ © 2016 Published by Elsevier Ltd.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).
E.J. Ciaccio et al. Informatics in Medicine Unlocked 6 (2017) 12–27

tive methods have been developed for signal analysis to distinguish the activation or activity
types and characteristics of VT [17,18] and for multielectrode record- 'body-surface' or 'body surface'
ing of these arrhythmias [19]. All of these methods can be applied to catheter
image analysis in two dimensions, with the main imaging modalities correlation
used to quantitatively characterize these arrhythmias being magnetic CT or 'computed tomography'
resonance imaging (MRI), computerized tomography (CT), positron ‘dominant frequency’
emission tomography (PET), and ultrasound. Doppler
There have been a number of seminal papers which have served to ECG or EKG or electrocardiogram
change the course of quantitative analysis and publication in the field of eigen or eigenvector or eigenvalue or ‘principal component’ or
arrhythmia research. For AF studies, it was first suggested that AF ‘principal components’
causes electrophysiological changes of the atrial myocardium that electroanatomic
could be responsible for the progressive nature of the arrhythmia in ‘ensemble averaging’ or ‘ensemble average’ or ‘signal averaging’ or
1995 [20]. The time course for electrical remodeling of the substrate in ‘signal average’
AF was shown definitively in 1996 [21]. Treatment of a focal source of entropy
AF using radiofrequency ablation was done in 1997 [22]. The large risk filter or filtered or filtering
of death for AF patients became of especial concern after the publica- fractals or fractal or chaos or nonlinear or Lyapunov
tion of the Framingham Heart Study [23]. The finding that ectopic Fourier or fourier
beats originating from foci in the pulmonary veins leading to AF, which fractionated or fractionation
responds to radiofrequency ablation, was published in 1998 [24]. The ‘frequency analysis’ or ‘power spectrum’
use of circumferential radiofrequency ablation of pulmonary vein ostia ‘f-wave’ or ‘F-wave’ or ‘f wave’ or ‘F wave’
to treat AF was published in 2000 [25]. Also in 2000, clinical outcomes ‘high-density’ or 'high density'
were found to be improved for patients with refractory AF by using a intracardiac
combination of AV node ablation and permanent pacing therapy [26]. intramural
A meta-analysis was done to show the efficacy of antiarrhythmic drugs lasso
versus radiofrequency ablation in 2009, with the finding that there are ‘linear prediction’ or ‘autoregressive moving average’ or ARMA
advantages to radiofrequency ablation [27]. In basic science, the mapping or map or maps
development of a new animal model of atrial fibrillation using chronic morphology
rapid atrial pacing was published in 1995 [28]. For VT studies, in 1995, MRI or ‘magnetic resonance’
epicardial mapping was achieved without complications [29]. In PET or 'positron emission tomography'
another study, published in 1997, electroanatomic mapping of the phase or phasic
heart was enabled without fluoroscopy [30]. In the same year, 1997, potential or voltage
the relationship between the thinnest infarct border zone and the quantitative or quantitation
isthmus of the reentrant circuit driving VT was shown [31]. Empirical regularity or organization or complexity
antitachycardia pacing for fast VT was noted to be more efficacious similarity
than shocks from ICDs in 2004 [32]. speckle
The above studies suggest that in the mid 1990's and 2000's, new ‘three-dimensional’ or 'three dimensional' or 3D
quantitative avenues and approaches for studying AF were developed, ‘time domain’ or ‘time series’
thus stimulating an interest in and possible increased volume of transform
research and published contrast on this topic. The development of a ‘two-dimensional’ or 'two dimensional' or 2D
number of curative therapies for VT during this same time period may ultrasound or ultrasonography
have somewhat reduced interest and published quantitative research wave or wavefront or wavebreak
on this topic. However, postinfarction VT treatment with catheter wavelet or wavelets
ablation is sometimes difficult and research to better characterize the
reentrant circuit is ongoing. In order to investigate these research Note that ‘US’ was not used as an abbreviation for ‘ultrasound’, as it
topics further, in this study, the MEDLINE search tool was used to is also used for other abbreviations in the medical literature.
identify trends in the biomedical literature for the quantitative analysis To search for a keyword in association with AF for example, the
of AF and VT. By identifying trends, it is possible to make inferences following syntax was used –
concerning the focus and direction of research in the field, so that the ‘atrial fibrillation’ and (‘time series’ or ‘time domain’).
state of the art may be improved. When a keyword consisted of more than one word, it was bracketed
by quotes, as shown above, which is required by the MEDLINE search
2. Method tool. Multiple words were bracketed by parentheses. The logical ‘or’
was used to combine the occurrence of several variants of a keyword.
In order to search for trends in quantitative arrhythmia research, the For the example above,
online MEDLINE database was used. The MEDLINE database works by
‘time series’ or ‘time domain’
determining the instance of a searched word in the title, abstract, and list of
keywords of a paper. It searches for peer-reviewed papers with instances in were used to determine any and all research papers containing
all journals included in the MEDLINE database. The tool was used to these keywords, which are two common ways to refer to the temporal
search over successive years from 1960 to 2013. The last year was selected analysis of the AF signal. The logical ‘and’ was used to determine those
as 2013 because there is a delay to finalize the database after each year. AF research papers in which at least one of the associated keywords appear
and VT were searched using the terms ‘atrial fibrillation’ and ‘ventricular in the literature with ‘atrial fibrillation’. The process was repeated using
tachycardia’, respectively. the same keywords in association with ‘ventricular tachycardia’.
The following list was developed in part from a MEDLINE search of MEDLINE performs the search based upon words found in the Title,
the most common quantitative terms associated with ‘atrial fibrillation’ Abstract, and Keywords sections of the article.
and ‘ventricular tachycardia’, shown below in alphabetical order. There The keywords for AF and VT were separately searched for the range in
are associated keywords (40 in all), and their variants, used for years 1960–2013 inclusive. The year 1960 was used as a starting point
identification purposes in this study: because it can be considered a watershed year for quantitative arrhythmia

13
E.J. Ciaccio et al.
studies – it was the first year in which cardiac pacing was used to prevent
VT in patients with complete heart block, and the first year in which AC
current was used for termination of VT [33]. Another major step which
took place in the 1960 's was to combine programmed electrical stimulation
of the myocardium with intracardiac activation recording [34], thus
ushering in a new era of quantitative cardiology.
Once data was obtained from MEDLINE for AF and VT, overall and
in association with each of the 40 keywords, it was copied into separate
text files. Using an in-house developed software program written in the
Fortran computer language (Intel Visual Fortran 64 Compiler XE, Intel
Corporation, Santa Clara, CA, 2015), the year of each publication was
determined for each of the 40 keywords in association with AF and VT.
The program automatically tabulated all papers published in the
biomedical literature for each year from 1960 to 2013. This was done
for all AF and VT studies in conjunction with the 40 associated
keywords, for studies found in the biomedical literature. The program
was also used to determine the number of AF and VT studies overall
that were published in the biomedical literature, and separately, how
many of the overall studies were published in the biomedical engineer-
ing and bioinformatics literature. This latter analysis was done by
determining, for each study, whether the journal in which it was
published belonged to the ‘Engineering, Biomedical’ or ‘Mathematical
& Computational Biology’ categories in the Journal Citation Reports
(Thomson Reuters, New York, 2015). The number of such studies was
again automatically tabulated per year by the software program.
The data for published studies in the biomedical literature overall
was then graphed as number of publications per year versus year from
1960 to 2013. Comparisons were shown by graphing the AF trace in
black with the VT trace in red. This procedure was also used to graph
the overall bioengineering and bioinformatics studies. Finally, the
association of keywords with AF and VT was graphed (40 graphs in
all). For perspective, each graph was redone by dividing the number of
studies per year of AF or VT associated with each keyword, by the
overall number of studies of AF or VT in the biomedical literature
(fractional association).
Linear regression analysis was then used to show trends in keyword
association for AF. The regression line onset, used to characterize AF
trends, was taken as the first year in which the associated keyword regularly
appeared when only a single trend was evident from 1960 to 2013. When
multiple fluctuations in the number of publications occurred over the
queried interval, the onset of the trend of greatest magnitude after any
initial fluctuation interval was selected. The linear regression line was
constructed using SigmaPlot (Ver. 13.0, Systat Software, Inc., San Jose, CA,
2014). The linear regression line is the best fit straight line through the data
points based on the least mean squares error. Year of first instance of the
associated keyword, linear regression coefficient, and trend in terms of the
slope of the regression line were tabulated for all queries. The linear
regression coefficient is the correlation between the data points and a
straight line. When all the data points reside on the regression line the
linear regression coefficient=1.0.
3. Results
A direct comparison of AF (black trace) and VT (red trace)
publications in the biomedical literature is shown in Fig. 1A. The
ordinate axis gives the total number of papers associated with each
keyword from 1960 to 2013. Very few AF and VT papers appeared in
the medical literature in 1960. However, the keyword ‘atrial fibrillation’
jumped in number of publications for the year 1963, and then
maintained an approximately constant value until 1978. Since 1978,
there has been a sharp increase in published biomedical research on AF
through 2013, when over 3500 articles were published per year. In
contrast, keyword ‘ventricular tachycardia’ showed a substantial in-
crease from 1975 to 1992, at which point it reached a peak of
approximately 600 peer-reviewed papers, whereupon the number of
papers published in subsequent years has been mostly unchanging
14
Informatics in Medicine Unlocked 6 (2017) 12–27
through 2013. In 1992, the peer-reviewed AF medical research over-
took published research on VT. From Fig. 1A, the rate of increase in the
number of published research papers per year pertaining to AF since
1992 has been approximately 135.60 based upon a linear regression
analysis (r2=0.949). In contrast, the rate of increase in the number of
published research papers per year pertaining to VT since 1992 has
been only 6.38 (r2=0.303).
In Fig. 1B–D are shown graphs which include traces of publications for
biomedical engineering and bioinformatics journals (denoted as Eng.).
From the traces of Fig. 1B–D, it is evident that bioengineering and
bioinformatics papers comprise approximately 10% of all biomedical
papers published on the topics of atrial fibrillation and ventricular
tachycardia. Substantial numbers of bioengineering and bioinformatics
papers on these topics began to appear in the year 1981 (Fig. 1B). A greater
number of atrial fibrillation as compared with ventricular tachycardia
bioengineering papers appeared in the literature after 1996 (Fig. 1B).
In Tables 1 and 2 are shown the results of linear regression analysis
for AF and VT studies, when ‘atrial fibrillation’ or ‘ventricular
tachycardia’ are shown in combination with the commonly associated
keywords. The columns note the 2013 value for number of publications
in a year, the year used at onset of the linear regression analysis
according to the presence of a trend, the linear regression coefficient,
and the slope of the regression line (trend). The listing is in order of
greatest to least 2013 AF publications. For ease of comparison, Table 1
shows keywords with a value of published studies in 2013 greater than
27, while Table 2 shows keywords with 2013 values less than 27. For
example, in Table 1, evident in the first row, there were 806 studies
published in which ‘atrial fibrillation’ and ‘catheter’ appeared in the
same study (based upon Abstract, Title, and/or keywords). The trend
for this association in the published literature began in 1973. The r2
value, which is the correlation of the yearly values to a straight line, was
0.74 for AF (1.0=maximum). The magnitude of the trend was 16.21,
meaning that on the average there was an increase of 16.21 publica-
tions per year since 1973. Shown in bold font for both AF and VT in
Tables 1 and 2 are trends of larger magnitude (greater slopes of the
regression line) which tend to be those keywords with the largest
number of publications in association with AF and VT in 2013. This
includes keywords ‘catheter’, ‘potential’, ‘ECG’, ‘activation’, ‘mapping’,
‘MRI’, and ‘wavefront’. Highly linear trends include keywords ‘map-
ping’ and ‘wavefront’ (r2=0.90) which are in association with VT.
Figs. 2–6 show the results of Tables 1 and 2 in graphical form, and
in the same order as is in the Tables. Black traces designate the number
of AF publications per year while red traces depict the number of VT
traces per year. In many of the graphs of Fig. 2, the VT studies in
association with keywords show a peak during the years 1990–1995,
then level off or slowly increase again (red traces). In general, AF
studies in association with keywords increased rapidly per year, with
some fluctuation (Fig. 2). Similar observations can be made for the
keywords depicted in Figs. 3–6, which had lesser publications in
association with AF in 2013. However, as the number of publications
per year of a particular keyword in association with VT or AF decreases,
the rapid increase in AF studies often associated with that keyword in
later years, and the peak in VT studies during the years 1990–1995,
when present, tend to fluctuate more, and are therefore likely to be less
reliable as an indicator of publishing characteristics (Figs. 3–6).
Many of the studies with the most publications in 2013 (i.e., in
Fig. 2) begin long upward trends in early years such as for keywords
‘ECG’ and ‘activation’ (about 1974), and ‘potential’, ‘mapping’, ‘CT’,
and ‘phase’ (1980's). Whereas, studies with keywords in association
with AF having the fewest publications in 2013 (Figs. 5–6) had very
recent onset of upward trends (late 1990's or 2000's) including
‘fractionated’, ‘high-density’, ‘speckle’, ‘dominant frequency’ and ‘elec-
troanatomic’ (Fig. 5) and ‘lasso’, ‘transform’ ‘similarity’, and ‘linear
prediction’ (Fig. 6).
Figs. 7–11 show the same data as Figs. 2–6 except that each trace is
scaled by the total number of biomedical publications (AF or VT) per
E.J. Ciaccio et al. Informatics in Medicine Unlocked 6 (2017) 12–27

Fig. 1. Overall relationship of atrial fibrillation (black trace) versus ventricular tachycardia (red trace) for published papers in the biomedical literature per year from 1960 to 2013.
Studies published in bioengineering and bioinformatics journals are shown in panels B–D. (For interpretation of the references to color in this figure legend, the reader is referred to the
web version of this article.)

year. Thus these are traces of individual keywords in association with
AF or VT normalized by all AF or VT, making the scale on the ordinate
axis fractional. The traces increase where the number of publications of
a particular keyword in association with AF or VT is fractionally greater
with respect to all AF or VT publications in the same years. In many of
the graphs of Figs. 7–11, the fractional traces of VT keywords are
greater than those of AF keywords. This means that there tend to be
more quantitative VT studies with respect to all VT studies published in
the biomedical literature, as compared with the number of quantitative
AF studies with respect to all AF studies published in the biomedical
literature. Therefore, although published VT studies in the biomedical
literature are lesser in number, they tend to be more quantitative than
AF studies. As for the absolute traces of Figs. 2–5, the fractional traces
of Figs. 7–11 tend to be more jagged, and likely less reliable, when
there are lesser published studies per year. Thus there is substantial
fluctuation in the fractional studies per year particularly evident in
Figs. 9–11. Some studies where trends are clearer, in terms of a smooth
curve, tend to show greater quantitative studies with respect to all

Table 1
Parameters, Larger 2013 Publications.

Term 2013 AF Trend AF r2AF Slope AF 2013 VT Trend VT r2VT Slope VT

Catheter 806 1973 0.74 16.21 170 1960 0.86 2.67

Potential 464 1960 0.68 6.49 222 1960 0.83 2.72
ECG 314 1960 0.78 4.03 169 1960 0.86 2.70
Activation 286 1960 0.76 4.14 81 1960 0.78 1.91
Mapping 182 1960 0.68 2.93 105 1973 0.90 2.14
MRI 166 1988 0.80 4.76 85 1984 0.65 1.85
CT 145 1980 0.74 3.37 24 1986 0.56 0.46
Phase 132 1960 0.65 1.6 22 1970 0.48 0.51
Correlation 117 1967 0.81 1.85 17 1975 0.19 0.28
Wavefront 114 1960 0.84 2.27 48 1975 0.90 1.09
Three-dimensional 104 1960 0.51 1.26 27 1986 0.82 0.83
Doppler 101 1960 0.85 1.94 9 1989 0.07 0.11
Ultrasound 54 1968 0.76 1.06 6 1984 0.44 0.16
Intracardiac 52 1969 0.81 1.07 17 1975 0.55 0.3
Regularity 51 1973 0.77 1.23 7 1979 0.23 0.13
Body surface 50 1960 0.55 1.08 33 1981 0.69 0.83
Fractionated 44 1989 0.64 2.22 8 1981 0.058 0.051
Two-dimensional 43 1979 0.54 0.6 5 1982 0.0027 −0.015
Morphology 36 1974 0.8 0.76 32 1979 0.5 0.62
Quantitative 28 1969 0.77 0.54 4 1983 0.00019 0.0033

15
E.J. Ciaccio et al. Informatics in Medicine Unlocked 6 (2017) 12–27

Table 2
Parameters, Lesser 2013 Publications.

Term 2013 AF Trend AF r2AF Slope AF 2013 VT Trend VT r2VT Slope VT

High-Density 26 1991 0.73 0.88 5 2002 0.58 0.36

Speckle 25 2006 0.68 2.53 2 2010 0.69 0.6
Dominant Frequency 23 1996 0.85 1.5 0 2000 0.0083 −0.03
Electroanatomic 23 1998 0.62 1.46 19 1999 0.85 1.09
Time Domain 20 1989 0.54 0.41 2 1987 0.074 −0.15
Fractals 15 1989 0.69 0.42 1 1993 0.014 0.026
Intramural 15 2006 0.5 1.1 9 1982 0.35 0.14
Entropy 9 1995 0.76 0.63 1 1994 0.027 −0.019
Fourier 7 1988 0.59 0.36 0 1982 0.079 −0.1
Frequency Analysis 7 1988 0.56 0.36 1 1985 0.28 −0.096
Filter 7 1988 0.35 0.33 3 1981 0.049 −0.12
Wavelet 7 1990 0.37 0.31 1 1993 0.00035 0.0045
Lasso 7 2002 0.02 0.11 0 2005 0.16 −0.04
Ensemble Average 5 1991 0.00043 −0.0052 0 1981 0.3 −0.19
F-wave 5 1987 0.24 0.11 0 – 0 0
Transform 5 1996 0.58 0.51 0 1987 0.09 −0.06
Eigen 4 2004 0.15 0.29 1 1995 0 0
Similarity 4 2002 0.42 0.31 0 2003 0.03 −0.059
Linear Prediction 3 2009 0.61 0.46 0 2000 0.087 −0.032
PET 1 1998 0.26 0.2 3 1991 0.36 0.16

– = not used.

studies in earlier years, including those for keywords ‘correlation’,
‘Doppler’, which have a peak for both AF and VT at approximately year
1990, and ‘intracardiac’, and ‘regularity’, peaking in 1980 (Fig. 8).
4. Discussion
4.1. Summary
In this study, the association of 40 quantitative keywords, typically
identified with AF and VT in the medical literature, were quantified.
For this purpose, the MEDLINE search tool was used. For each year
from 1960 to 2013, the overall number of research papers in which
each keyword appeared with AF and VT was determined. A comparison
of papers per year in which AF versus VT appear is provided in Fig. 1A.
AF and VT papers were present in the biomedical research literature in
1960, the first year shown in the graph, but were uncommon. Early
after 1960 there was a sharper increase in VT papers in the literature.
However, the number of AF papers published per year overtook VT
papers in 1993, and for AF began to increase rapidly. This may have
resulted from a number of seminal quantitative curative methods
developed for VT and published in the literature about 1995–2000,
as compared to quantitative AF studies published at the same time
which may have spurred additional research studies. In 2013, there
were approximately five times as many quantitative AF papers as VT
papers appearing in the medical literature. The overall number of these
papers published in the bioengineering and bioinformatics literature
was approximately 10% that of all studies published in the biomedical
literature (Fig. 1B–D).
Results from the study were also graphed as the number of AF and
VT papers containing the keyword versus year from 1960 to 2013 (40
keywords) as shown in Figs. 2–6. These graphs were ordered in terms
of the association of the keyword with AF in 2013, from highest to
lowest association. All of the data was further quantified using linear
regression analysis (Tables 1 and 2). The linear regression parameters
that were determined were the onset year for the linear regression line,
the slope of the regression line, which is the trend in the association,
and the regression coefficient or r2 value, which is the degree of
linearity of the trend. Strong linear trends are those with steep slopes
and r2 values closer to 1.0, as can be observed in Tables 1 and 2.
Another reason for the early peak of quantitative VT studies may have
been the ability of ventricular tachycardia to first be initiated and
terminated by critically-timed premature ventricular beats in 1971
[33]. However, a number of quantitative keywords in association with
VT tended to peak in the early1990's, notably ‘activation’ (Fig. 2),
‘correlation’ (Fig. 3), ‘quantitative’ (Fig. 4), ‘Time Domain’, ‘Fourier’,
and ‘Filter’ (Fig. 5) and ‘Ensemble Average’ (Fig. 6). Perhaps, this may
have been in part due to the significant impact of ablation as a therapy
to stop VT, which began in 1987 [35], resulting in a lesser rate of
overall increase in published studies on VT as a clinical and quantita-
tive research topic as compared to AF (Fig. 1A). By comparison, the
mechanisms for AF, and the best treatment procedures, are still mostly
unknown. Therefore, most quantitative keywords in association with
AF were still increasing in number of publications per year as of 2013
(Figs. 2–6). However, there appears to be a decrease since the late
2000's for some keywords including ‘body surface’ (Fig. 3),
‘Fractionated’ and ‘Electroanatomic’ (Fig. 4) ‘Entropy’, ‘Frequency
Analysis’, ‘Filter’, and ‘Wavelet’ (Fig. 5), and ‘lasso’, ‘Transform’,
‘Eigen’, and ‘Similarity’ (Fig. 6). When the data was graphed as a
fraction of all studies (Figs. 7–11), quantitative keywords tended to
appear fractionally more frequently in VT publications as compared
with AF publications.
4.2. Trends in associated keywords
From the Tables, several keywords showed significant upswings in
the published literature in recent years: 'catheter', 'potential', 'ECG',
'activation', 'mapping', and 'MRI'. Both AF and VT studies often involve
catheter mapping and catheter ablation. The voltage or potential level
of the electrogram is important to gauge the viability of the substrate
for conducting the electrical activation wavefront. And magnetic
resonance imaging or MRI is becoming an important tool for determin-
ing the structural characteristics of the myocardium, which can readily
effect the activation wavefront [36]. Thus the increasing inclusion of
these keywords in both the quantitative and clinical research literature
is likely to be helpful for further understanding of the mechanisms and
best treatment options for both AF and VT.
4.3. Limitations
The study was done in part using the MEDLINE search tool. Other tools
that are available for searching the medical research literature such as
PubMed and Science Citation Index were not used. Use of other search

16
E.J. Ciaccio et al. Informatics in Medicine Unlocked 6 (2017) 12–27

Fig. 2. Keywords commonly associated with atrial fibrillation and ventricular tachycardia (range of association with atrial fibrillation 806 – 132 studies per year in 2013). Shown are the
trend from 1960 to 2013 (atrial fibrillation, black trace, ventricular tachycardia, red trace). (For interpretation of the references to color in this figure legend, the reader is referred to the
web version of this article.)

17
E.J. Ciaccio et al. Informatics in Medicine Unlocked 6 (2017) 12–27

Fig. 3. Keywords commonly associated with atrial fibrillation and ventricular tachycardia (range of association with AF 117 – 50 studies per year in 2013). Shown are the trend from
1960 to 2013 (atrial fibrillation, black trace, ventricular tachycardia, red trace).(For interpretation of the references to color in this figure legend, the reader is referred to the web version
of this article.)

18
E.J. Ciaccio et al. Informatics in Medicine Unlocked 6 (2017) 12–27

Fig. 4. Keywords moderately associated with atrial fibrillation and ventricular tachycardia (range of association with atrial fibrillation 44 – 23 studies per year in 2013). Shown are the
trend from 1960 to 2013 (AF, black trace, ventricular tachycardia, red trace). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of
this article.)

19
E.J. Ciaccio et al. Informatics in Medicine Unlocked 6 (2017) 12–27

Fig. 5. Keywords moderately associated with atrial fibrillation and ventricular tachycardia (range of association with atrial fibrillation 20 – 7 studies per year in 2013). Shown are the
trends from 1960 to 2013 (atrial fibrillation, black trace, ventricular tachycardia, red trace). (For interpretation of the references to color in this figure legend, the reader is referred to the
web version of this article.)

20
E.J. Ciaccio et al. Informatics in Medicine Unlocked 6 (2017) 12–27

Fig. 6. Keywords moderately associated with atrial fibrillation and ventricular tachycardia (range of association with atrial fibrillation 7 – 1 studies per year in 2013). Shown are the
trends from 1960 to 2013 (atrial fibrillation, black trace, ventricular tachycardia, red trace). (For interpretation of the references to color in this figure legend, the reader is referred to the
web version of this article.)

tools may generate slightly different results, as some journals and articles
not reflected in one search tool may appear in another. The data was
generated from year-to-year. In early years from 1960, there were often few
instances of each keyword in association with AF; thus fluctuation in the
prevalence of these keywords tended to occur early in the studied time
interval. Fluctuations of associated quantitative keywords with VT were
common due to the lesser frequency of such studies. The results of all
keywords with lesser published studies per year, as in Table 2, should be
considered less reliable. The linear regression models used for parameter-
ization of the data as shown in Tables 1 and 2 are estimates and
approximations. Use of nonlinear models would better fit many of the
curves in Figs. 2–6. The 40 quantitative associated keywords were selected

21
E.J. Ciaccio et al. Informatics in Medicine Unlocked 6 (2017) 12–27

Fig. 7. Keywords commonly associated with atrial fibrillation and ventricular tachycardia (range of association with atrial fibrillation 806 – 132 studies per year in 2013). Values for AF
and VT traces are scaled by the total number of publications for AF and VT, respectively from 1960 to 2013 (atrial fibrillation, black trace, ventricular tachycardia, red trace). (For
interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

as those typically identified with AF and VT, but the list is not exhaustive, would be useful for further study of publication trends [37]. For verification,
and other keywords may be associated with these arrhythmias as well. the work should be repeated in a larger study in which all keywords
Therefore, inclusion of other associated keywords from recent reviews appearing in the medical literature with AF and VT are quantified.

22
E.J. Ciaccio et al. Informatics in Medicine Unlocked 6 (2017) 12–27

Fig. 8. Keywords commonly associated with atrial fibrillation and ventricular tachycardia (range of association with AF 117 – 50 studies per year in 2013). Values for AF and VT traces
are scaled by the total number of publications for AF and VT, respectively from 1960 to 2013 (atrial fibrillation, black trace, ventricular tachycardia, red trace). (For interpretation of the
references to color in this figure legend, the reader is referred to the web version of this article.)

23
E.J. Ciaccio et al. Informatics in Medicine Unlocked 6 (2017) 12–27

Fig. 9. Keywords moderately associated with atrial fibrillation and ventricular tachycardia (range of association with atrial fibrillation 44 – 23 studies per year in 2013). Values for AF
and VT traces are scaled by the total number of publications for AF and VT, respectively from 1960 to 2013 (AF, black trace, ventricular tachycardia, red trace). (For interpretation of the
references to color in this figure legend, the reader is referred to the web version of this article.)

5. Conclusions relationship of quantitative keywords typically identified with AF and

VT in the medical literature. A technique that was devised to quantify
The MEDLINE search technique, in conjunction with an in-house first instance, association, and trends of keywords with AF and VT in
developed software program, was shown to be helpful to determine the literature studies, based on the presence of these words in title,

24
E.J. Ciaccio et al. Informatics in Medicine Unlocked 6 (2017) 12–27

Fig. 10. Keywords moderately associated with atrial fibrillation and ventricular tachycardia (range of association with atrial fibrillation 20 – 7 studies per year in 2013). Values for AF
and VT traces are scaled by the total number of publications for AF and VT, respectively from 1960 to 2013 (atrial fibrillation, black trace, ventricular tachycardia, red trace). (For
interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

abstract, or keywords of each entry, was successful for characterization certain keywords. Although AF and VT are different arrhythmias, both
of the AF and VT relationships. From the graphs in Figs. 1–6, and from have increased in the published medical literature since 1960, but AF
the linear regression parameters of Tables 1 and 2, it is evident that studies have overtaken VF studies since 1993, and are predominant as
significant associations and trends exist for AF and VT with respect to compared with VT studies in recent years.

25
E.J. Ciaccio et al. Informatics in Medicine Unlocked 6 (2017) 12–27

Fig. 11. Keywords moderately associated with atrial fibrillation and ventricular tachycardia (range of association with atrial fibrillation 7 – 1 studies per year in 2013). Values for AF
and VT traces are scaled by the total number of publications for AF and VT, respectively from 1960 to 2013 (atrial fibrillation, black trace, ventricular tachycardia, red trace). (For
interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

References
[1] January CT, Wann LS, Alpert JS, Calkins H, Cleveland JC, Jr, Cigarroa JE, Conti
JB, Ellinor PT, Ezekowitz MD, Field ME, Murray KT, Sacco RL, Stevenson WG,
Tchou PJ, Tracy CM, Yancy CW. AHA/ACC/HRS Guideline for the management of
patients with atrial fibrillation: executive summary. A report of the american college
of cardiology/american heart association task force on practice guidelines and the
Heart Rhythm Society. Circulation 2014;2014, [Epub ahead of print].
[2] Zipes DP, Camm AJ, Borggrefe M, Buxton AE, Chaitman B, Fromer M, Gregoratos

26
E.J. Ciaccio et al.
G, Klein G, Moss AJ, Myerburg RJ, Priori SG, Quinones MA, Roden DM, Silka MJ,
Tracy C, Blanc JJ, Budaj A, Dean V, Deckers JW, Despres C, Dickstein K, Lekakis J,
McGregor K, Metra M, Morais J, Osterspey A, Tamargo JL, Zamorano JL, Smith
SC, Jr, Jacobs AK, Adams CD, Antman EM, Anderson JL, Hunt SA, Halperin JL,
Nishimura R, Ornato JP, Page RL, Riegel B. American college of cardiology/
American heart association task force; European Society of Cardiology Committee
for practice guidelines; European heart Rhythm Association and the heart Rhythm
society. ACC/AHA/ESC 2006 guidelines for management of patients with ventri-
cular arrhythmias and the prevention of sudden cardiac death–executive summary:
a report of the American College of Cardiology/American heart Association task
force and the European Society of Cardiology Committee for practice guidelines
(writing Committee to develop Guidelines for Management of Patients with
Ventricular Arrhythmias and the Prevention of Sudden Cardiac Death) developed in
collaboration with the European heart Rhythm Association and the heart Rhythm
Society. Eur Heart J 2006;27(17):2099–140.
[3] European Heart Rhythm Association , European Association for Cardio-Thoracic
Surgery , Camm AJ, Kirchhof P, Lip GY, Schotten U, Savelieva I, Ernst S, Van
Gelder IC, Al-Attar N, Hindricks G, Prendergast B, Heidbuchel H, Alfieri O,
Angelini A, Atar D, Colonna P, De Caterina R, De Sutter J, Goette A, Gorenek B,
Heldal M, Hohloser SH, Kolh P, Le Heuzey JY, Ponikowski P, Rutten FH.
Guidelines for the management of atrial fibrillation. The task force for the
management of atrial fibrillation of the European Society of Cardiology (ESC). Eur
Heart J 2010;31:2369–429.
[4] Ropella KM, Baerman JM, Sahakian AV, Swiryn S. Differentiation of ventricular
tachyarrhythmias. Circulation 1990;82(6):2035–43.
[5] Nattel S, Burstein B, Dobrev D. Atrial remodeling and atrial fibrillation: mechan-
isms and implications. Circ Arrhythm Electrophysiol 2008;1:62–73.
[6] Bunch TJ, Crandall BG, Weiss JP, May HT, Bair TL, Osborn JS, Anderson JL,
Muhlestein JB, Horne BD, Lappe DL, Day JD. Patients treated with catheter
ablation for atrial fibrillation have long-term rates of death, stroke, and dementia
similar to patients without atrial fibrillation. J Cardiovasc Electrophysiol
2011;22:839–45.
[7] Kottkamp H, Tanner H, Kobza R, Schirdewahn P, Dorszewski A, Gerds-Li JH,
Carbucicchio C, Piorkowski C, Hindricks G. Time courses and quantitative analysis
of atrial fibrillation episode number and duration after circular plus linear left atrial
lesions: trigger elimination or substrate modification: early or delayed cure?. J Am
Coll Cardiol 2004;44:869–77.
[8] Choudhry NK, Zagorski B, Avorn J, Levin R, Sykora K, Laupacis A, Mamdani M.
Comparison of the impact of the Atrial Fibrillation Follow-Up Investigation of
Rhythm Management trial on prescribing patterns: a time-series analysis. Ann
Pharmacother 2008;42:1563–72.
[9] Ciaccio EJ, Coromilas J, Wit AL, Garan H. Onset dynamics of ventricular
tachyarrhythmias as measured by dominant frequency. Heart Rhythm
2011;8(4):615–23.
[10] Corino VD, Mainardi LT, Stridh M, Sörnmo L. Improved time–frequency analysis
of atrial fibrillation signals using spectral modeling. IEEE Trans Biomed Eng
2008;55:2723–30.
[11] Peters RM. The fractal dimension of atrial fibrillation: a new method to predict left
atrial dimension from the surface electrocardiogram. Cardiology 1999;92:17–20.
[12] Lake DE, Moorman JR. Accurate estimation of entropy in very short physiological
time series: the problem of atrial fibrillation detection in implanted ventricular
devices. Am J Physiol Heart Circ Physiol 2011;300:H319–H325.
[13] Alcaraz R, Rieta JJ. Central tendency measure and wavelet transform combined in
the non-invasive analysis of atrial fibrillation recordings. BioMed Eng OnLine
2012;11:46.
[14] Everett TH, 4th, Akar JG, Kok LC, Moorman JR, Haines DE. Use of global atrial
fibrillation organization to optimize the success of burst pace termination. J Am
Coll Cardiol 2002;40:1831–40.
[15] Zhao J, Butters TD, Zhang H, Pullan AJ, LeGrice IJ, Sands GB, Smaill BH. An
image-based model of atrial muscular architecture: effects of structural anisotropy
on electrical activation. Circ Arrhythm Electrophysiol 2012;5:361–70.
[16] Pappone C, Santinelli V. Multielectrode basket catheter: a new tool for curing atrial
fibrillation?. Heart Rhythm 2006;3:385–6.
[17] Throne RD, Windle JR, Easley AR, Jr, Olshansky B, Wilber D. Scatter diagram
analysis: a new technique for discriminating ventricular tachyarrhythmias. Pacing
Clin Electrophysiol 1994;17(7):1267–75.
27
Informatics in Medicine Unlocked 6 (2017) 12–27
[18] Gold MR, Theuns DA, Knight BP, Sturdivant JL, Sanghera R, Ellenbogen KA, Wood
MA, Burke MC. Head-to-head comparison of arrhythmia discrimination perfor-
mance of subcutaneous and transvenous ICD arrhythmia detection algorithms: the
start study. J Cardiovasc Electrophysiol 2012;23(4):359–66.
[19] Greenspon AJ, Hsu SS, Datorre S. Successful radiofrequency catheter ablation of
sustained ventricular tachycardia postmyocardial infarction in man guided by a
multielectrode ‘basket’ catheter. J Cardiovasc Electro 1997;8(5):565–70.
[20] Wijffels MC, Kirchhof CJ, Dorland R, Allessie MA. Atrial fibrillation begets atrial
fibrillation. A study in awake chronically instrumented goats. Circulation
1995;92:1954–68.
[21] Goette A, Honeycutt C, Langberg JJ. Electrical remodeling in atrial fibrillation.
Time course and mechanisms. Circulation 1996;94:2968–74.
[22] Jaïs P, Haïssaguerre M, Shah DC, Chouairi S, Gencel L, Hocini M, Clémenty J. A
focal source of atrial fibrillation treated by discrete radiofrequency ablation.
Circulation 1997;95:572–6.
[23] Benjamin EJ, Wolf PA, D'Agostino RB, Silbershatz H, Kannel WB, Levy D. Impact
of Atrial Fibrillation on the Risk of Death: the Framingham Heart Study.
Circulation 1998;98:946–52.
[24] Haïssaguerre M, Jaïs P, Shah DC, Takahashi A, Hocini M, Quiniou G, Garrigue S,
Le Mouroux A, Le Métayer P, Clémenty J. Spontaneous initiation of atrial
fibrillation by ectopic beats originating in the pulmonary veins. N Engl J Med
1998;339:659–66.
[25] Pappone C, Rosanio S, Oreto G, Tocchi M, Gugliotta F, Vicedomini G, Salvati A,
Dicandia C, Mazzone P, Santinelli V, Gulletta S, Chierchia S. Circumferential
radiofrequency ablation of pulmonary vein ostia a new anatomic approach for
curing atrial fibrillation. Circulation 2000;102:2619–28.
[26] Wood MA, Brown-Mahoney C, Kay GN, Ellenbogen KA. Clinical outcomes after
ablation and pacing therapy for atrial fibrillation: a meta-analysis. Circulation
2000;101:1138–44.
[27] Calkins H, Reynolds MR, Spector P, Sondhi M, Xu Y, Martin A, Williams CJ, Sledge
I. Treatment of atrial fibrillation with antiarrhythmic drugs or radiofrequency
ablation: two systematic literature reviews and meta-analyses. Circ Arrhythm
Electro 2009;2:349–61.
[28] Morillo CA, Klein GJ, Jones DL, Guiraudon CM. Chronic rapid atrial pacing.
Structural, functional, and electrophysiological characteristics of a new model of
sustained atrial fibrillation. Circulation 1995;91:1588–95.
[29] Sosa E, Scanavacca M, d'Avila A, Pilleggi F. A new technique to perform epicardial
mapping in the electrophysiology laboratory. J Cardiovasc Electrophysiol
1996;7:531–6.
[30] Gepstein L, Hayam G, Ben-Haim SA. A novel method for nonfluoroscopic catheter-
based electroanatomical mapping of the heart. Vitr vivo Accuracy Results Circ
1997;95:1611–22.
[31] Peters NS, Coromilas J, Severs NJ, Wit AL. Disturbed connexin43 gap junction
distribution correlates with the location of reentrant circuits in the epicardial
border zone of healing canine infarcts that cause ventricular tachycardia.
Circulation 1997;95:988–96.
[32] Wathen MS, DeGroot PJ, Sweeney MO, Stark AJ, Otterness MF, Adkisson WO,
Canby RC, Khalighi K, Machado C, Rubenstein DS, Volosin KJ. PainFREE Rx II
Investigators. prospective randomized multicenter trial of empirical antitachycar-
dia pacing versus shocks for spontaneous rapid ventricular tachycardia in patients
with implantable cardioverter-defibrillators: pacing fast ventricular tachycardia
reduces shock therapies (PainFREE Rx II) trial results. Circulation
2004;110:2591–6.
[33] McGovern B, Schoenfeld MH, Ruskin JN, Garan H, Yurchak PM. Ventricular
tachycardia: Historical perspective. Pacing Clin Electrophysiol 1986;9:449–62.
[34] Wellens HJJ. Cardiac arrhythmias: the quest for a cure. A Historical perspective. J
Am Coll Cardiol 2004;44:1155–63.
[35] Budde TH, Breithardt G, Borggreve M, Podczeck A, Langwasser J. Erste
Erfahrungen mit der hochfrequenzstrom ablation der AV leitungssystems beim
menschen. Z Kardiol 1987;76:204–10.
[36] Ciaccio EJ, Coromilas J, Wit AL, Peters NS, Garan H. Formation of functional
conduction block during the onset of reentrant ventricular tachycardia. Circ
Arrhythm Electro 2016;9(12), [in press] Dec, pii: e004462..
[37] Ravelli F, Masè M. Computational mapping in atrial fibrillation: how the integra-
tion of signal-derived maps may guide the localization of critical sources. Europace
2014;16:714–23.