411 / 1096
S396
ESTRO 36
_______________________________________________________________________________________________
this linked information to evaluate whether or not trials
have been published in a PRJ.
Results
Of 3479 oncology trials, 2551 (73%) did not deposit a
summary result in the registry; only 1096 (32%) had a peer-
reviewed publication of their results indexed; only 1596
(46%) had either deposit a summary result in the registry,
or published their results in a PRJ, or both. Of 1458
molecular oncology trials, 990 (68%) did not deposit a
summary result in the registry; only 514 (35%) had a peer-
reviewed publication of their results indexed; only 734
(50%) had either deposit a summary result in the registry,
or published their results in a PRJ, or both. Of 483
radiation oncology trials, 414 (85.7%) did not deposit a
summary result in the registry; only 154 (32%) had a peer-
reviewed publication of their results indexed; only 227
(47%) had either deposit a summary result in the registry,
or published their results in a PRJ, or both.
Conclusion
Our results show that most trials (70%-80%) in oncology did
not report the results in the registry, even though they
have been required to do so. More than half of these trials
might not have been published in the biomedical
literature. Molecular oncology and radiation oncology are
no exception to the results found in oncology. Our
evidence in oncology is therefore distorted in important
ways and this might lead, to say the least, to large
inefficiencies in our health care system.
PO-0751 Uptake of a novel interactive 3D web-based
contouring atlas among the radiation oncology
community
E. Gillespie
1
, N. Panjwani
1
, P. Sanghvi
1
, J. Murphy
1
1
University of California San Diego, Radiation Medicine
and Applied Sciences, La Jolla, USA
Purpose or Objective
In the era of highly conformal treatment techniques, the
delivery of safe and effective radiation therapy
increasingly relies on accurate target delineation. Current
contouring resources are fragmented and cumbersome to
use at the point of care. We created a free interactive 3D
web-based atlas called eContour (
www.eContour.org),
which displays best available evidence to guide contour
delineation. This study reports on user characteristics,
frequency of use, and frequently viewed cases during the
first 6 months of dissemination.
Material and Methods
To track individual user data and collect user feedback,
visitors to the eContour website are required to register
with their email address, profession and hospital
affiliation, while the IP address of the computer from
which the website is accessed is used to track the user’s
geographic location. Google Analytics and Mixpanel were
employed to track the number and duration of each page
view, which was linked to the unique registered user.
Descriptive statistics were reported, including frequency
of repeat use (primary endpoint) defined as the ratio of
users that accessed the website on two different days
divided by the total number of registered users. Users in
the top quintile were further characterized by profession
and geographic location.
Results
eContour has 2,616 registered users, of which most (60%)
are radiation oncologists (1,092 practicing physicians and
459 residents). Other users include dosimetrists (16%),
physicists (9%), radiation therapists (5%), and medical
students (5%). Registered users represent 81 countries,
with the majority (56%) of users located in the US (see
Figure for map). Overall rate of repeat use was 49%, with
residents most likely to return to the site (63%) and
physicists least likely (36%). Repeat users visited the site
on up to 58 different days during the dissemination period,
with residents visiting a mean 6.6 different days and
practicing physicians visiting a mean 5.4 different days.
Users in the top quintile were primarily physicians (69%)
and from the US (67%). Of 33 posted cases in H&N, GI,
GYN, GU and lymphoma, the most frequently viewed
disease site was H&N (8,171 case views) followed by GI
(4,665 case views) and the most frequently viewed cases
were nasopharynx, pre-op rectal, pre-op esophageal,
anal, post-op endometrial, and intact prostate cancer,
each with over 1,000 page views. Users viewed each case
page
for
an
average
of
3.7
minutes.
Conclusion
eContour has a high rate of repeat use especially among
radiation oncologists in the US. Users are most frequently
looking for help contouring GI and H&N cancers, and they
spend fewer than 4 minutes accessing information within
a single case. These preliminary data suggest that
eContour is a resource that fills a need among radiation
oncology professionals. By providing users with updated
contouring guidelines at the point of care, eContour has
the potential to improve contour accuracy and ultimately
impact quality of radiation delivery.
Poster: Clinical track: Other
PO-0752 Birth outcomes in female cancer patients
received radiotherapy: a nationwide population-based
study
W.H. Kao
1
, J.H. Hong
1,2
, C.C. Wang
1,2
, Y.J. Chiang
1
, C.F.
Kuo
3
1
Chang Gung Memorial Hospital, Department of
Radiation Oncology, Taoyuan, Taiwan
2
College of Medicine- Chang Gung University,
Department of Medical Imaging and Radiological
Sciences, Taoyuan City, Taiwan
3
Chang Gung Memorial Hospital, Department of
Rheumatology- Allergy and Immunology-, Taoyuan,
Taiwan
Purpose or Objective
For young female cancer survivors who ever received
radiotherapy, adverse pregnant outcomes are serious
questions not only for themselves but also for their babies.
The purpose of this study was to estimate the risks of
adverse foetal-neonatal outcomes in female cancer
patients received radiotherapy (RT) compared with
women without malignancies.
Material and Methods
We identified 2,350,335 singleton pregnancies using
Taiwan National Health Insurance Database and Taiwan
Birth Registry between 2001 and 2012, of which 607
pregnancies were in female cancer patients with RT. Odds
ratios (ORs) and 95% confidence intervals (CIs) for foetal-
neonatal outcomes were estimated using generalized
estimating equation model adjusted by maternal age,
income, occupation, Charlson comorbidity index,
urbanization, infant sex and birth of year.
Results
From 2001 to 2012, the mean age at pregnancy of female
cancer patients received radiotherapy was 33.5 years old.
There were no significant increasing risks with an adjusted
OR (95% CIs) of 0.50 (0.21-1.21) for stillbirth, 0.75 (0.55-