WORKFLOW
EFFICIENCIES IN ULTRASOUND
You are listening to ReachMD, The Channel for Medical
Professionals. Welcome to advances in medical imaging, a program discussing
the latest innovations in clinical radiology and imaging technologies. This
edition of advances in medical imaging is sponsored by Siemens Ultrasound, the
only provider of knowledge-based workflow applications that improve
efficiencies and generate consistent results. Siemens, answers for life.
Intelligent uses of artificial intelligence to improve
diagnostic imaging services. With me today is Dr. Raley Lazebnik. Dr.
Lazebnik is a radiologist with a PhD in biomedical engineering. He is the
global radiology segment manager for the Siemens Healthcare Ultrasound Business
Unit. Today we are discussing workflow efficiencies in ultrasound.
DR. JASON BIRNHOLZ:
Hello, Raley, thank you for joining us.
DR. ROEE LAZEBNIK:
It's my pleasure to be here, Jason. Thank you.
DR. JASON BIRNHOLZ:
Well, I wonder if we should think of you as a 21st century
medical efficiency expert.
DR. ROEE LAZEBNIK:
That would be nice, yes.
DR. JASON BIRNHOLZ:
Why, you have a lot of tools at your disposal to do it in a
way that business people were not able to do it nearly 50 years ago.
DR. ROEE LAZEBNIK:
I certainly think so. I think one of the benefits from
medical imaging today not only stems from the actual imaging hardware, but also
from the manipulation of the clinical data along the way. We see many of the
efficiencies coming from our ability to now extract much more clinical
information from the same exam.
DR. JASON BIRNHOLZ:
Well, we are into the information age, aren’t we?
DR. ROEE LAZEBNIK:
We certainly are.
DR. JASON BIRNHOLZ:
Well, lets get down to some clinical discussions. There are
a lot of similarities between the patient visiting his/her primary physician’s
care office or going to an imaging facility for a diagnostic procedure like
ultrasound, there is some fundamentals and analyzing workflow that you consider
before you even begin to look into ways to stiff things up?
DR. ROEE LAZEBNIK:
Absolutely. We think about maximizing workflow in terms of
increasing the amount of information density obtained from a clinical study, that
is not just the number of images, but the actual clinical information derived
from those images. At the same time, we would like to lower the total exam
time, that is the time required for acquisition, analysis of the data, as well
as the annotation and reporting associated with it. So when we design new
technologies to maximize information density, we are very cognizant of the need
to minimize acquisition time, decrease analysis time, and at the same time make
the reporting much easier.
DR. JASON BIRNHOLZ:
Ok, well, lets take a hypothetical patient and since there
are probably more obstetrical ultrasounds done than any other kind of
ultrasound worldwide daily basis, lets make this hypothetical person pregnant,
a woman who is, lets say, in the second trimester of pregnancy and she is
referred by a physician to an ultrasound facility with your latest and best
workflow improvements, your best scanning equipment. Can you walk the patient
through her side of the visit?
DR. ROEE LAZEBNIK:
Absolutely. I think these days the woman would enter the
ultrasound suite just as she has in the past, but once situated the sonographer
or the sonologist – indicates that the scanning physician, would perform the
exam using a volume transducer, that is rather than trying to obtain
independent 2-dimensional views from various directions, the woman is going to
experience a much shorter exam because now the transducer placed on her is able
to acquire an entire volume; for instance, of the fetus almost instantaneously
rather than the physician having to manipulate her, position her, and sweep the
transducer manually across her. I think that that’s a much more convenient
experience for the patient in this case.
DR. JASON BIRNHOLZ:
Well, when that’s done, can the patient be let go or should
she stay until more is done?
DR. ROEE LAZEBNIK:
I think that’s an excellent point. In fact the patient in
many cases can be let go at that point particularly for non-emergency related cases
that is for standard biometric measurements as well as for well visits that
simply document the growth of the fetus and so forth. Certainly the patient
can leave at this point because a comprehensive volume of data has already been
acquired and any view can be retrospectively extracted from that patient, either
on the ultrasound system itself by the sonographer for instance or even hours
later at an off-line workstation.
DR. JASON BIRNHOLZ:
Okay, so on a conventional exam, patient comes in and
whoever the operator is, is perhaps spending 30 or 40 minutes thereabout
acquiring lots of particular images or according to some protocol may be saving
video clips, etc. If you change things to volume acquisition, what’s the
change in time?
DR. ROEE LAZEBNIK:
You know, that certainly depends on the specifics of the
exam, but I think in terms of time savings, results are fairly dramatic. In
some cases, you might be looking at even an order of magnitude decrease
particularly for patients where it would have been difficult to obtain some
views due to fetal posturing for instance that is the specific orientation of
the fetus is no longer an issue if we are able to retrospectively extract a
view we didn’t acquire directly.
DR. JASON BIRNHOLZ:
Okay, and actually I have heard from some of my colleagues
that you know times go down to may be 4 or 5 minutes for exam.
DR. ROEE LAZEBNIK:
I think that would be reasonable in some cases absolutely.
DR. JASON BIRNHOLZ:
Well, I mean, there are a lot of advantages to that on the
patient’s side that you have minimal waiting time and you can probably get to
something close like on-demand scheduling because on the facility side, we are
able to do many more patients per day with the same equipment and the same
personnel.
DR. ROEE LAZEBNIK:
Absolutely, and I think that the other aspect of this is
that the exams can be much more standard across different users. So in the
past there has been significantly more skill required to obtain a specific
view. Today, the skill is more in the manipulation of the volume rather than
in its acquisition, which can be standardized across the users much more
effectively.
DR. JASON BIRNHOLZ:
Now, that’s a really important point because one of the
things that is always distinguished ultrasound is the requirements on the
knowledge and experience of the operator whoever that may be, a technologist or
a physician in that rest of the radiology, you can perform the imaging
functions without having to interpret or understand the images particularly
well. You just need to be able to decide if the image is satisfactory or not
and so that’s a very very big difference to decrease the requirement for a
great deal of skill in acquiring the images.
DR. ROEE LAZEBNIK:
Absolutely. I also think that this allows the clinician to
spend more time with the patient as opposed to with the imaging hardware that
is when the burden is taken off the physician to obtain certain views for
documentation purposes. Once the scan is initiated, I think the encounter can
be much more personal for the patient. The physician is much more free to talk
with the patient and answer questions during the exam and so forth.
DR. JASON BIRNHOLZ:
And acquire more history that helps the diagnosis as well.
DR. ROEE LAZEBNIK:
Absolutely. I think overall we are able through this to
increase the density of clinical information. The clinician is able to extract
from the same encounter both by interacting with the patient more effectively
and at the same time getting the diagnostic views quicker and more efficiently.
DR. JASON BIRNHOLZ:
If you are just joining us, you are listening to advances
in medical imaging on ReachMD, The Channel for Medical Professionals. I am Dr.
Jason Birnholz and I am speaking with Dr. Roee Lazebnik. We are discussing
workflow efficiencies in ultrasound.
Well, Roee, lets get to
what is the sort of under the hood of all this, tell us what happens on the
radiology side after a block of data is acquired?
DR. ROEE LAZEBNIK:
It's quite interesting. In the past, once data was
acquired, most of the analysis of that imaged data was manual observation of
the image and may be some manually placed annotations on the image. These days
we are able to automate a great deal of measurements using artificial
intelligence technology, for instance, in the case of fetal biometrics, if a
particular image demonstrates a fetal femur, the ultrasound machine can now
automatically identify that it is looking at a femur and measure that femur
automatically, transfer that measurement into the report, and in conjunction
with other biometric measurements perform different obstetrical calculations
again with minimal user intervention.
DR. JASON BIRNHOLZ:
But your artificial intelligence research has really gone
far beyond that, hasn’t it, actually recognizing patterns identifying organs or
may be even pathology?
DR. ROEE LAZEBNIK:
Well, I think, we are not quite <_____> in terms of
automatically recognizing pathology, although I think that’s certainly down the
line. I think at this point in time we have made great advances in terms of
automating some of the more routine aspects of ultrasound. The way we have
been able to do that is by transitioning our image analyzing strategies from
thinking like a computer to thinking like a radiologist or an obstetrician in
this case. We found that humans tend to look at things in a very probability
based way that is when they look at an image, they are able to determine, which
part of an image corresponds to the anatomy that they are looking for and which
parts of the image are clearly not relevant diagnostically. We are able to
essentially train the imaging software to perform that same type of analysis by
feeding it a very large database of human analyzed cases that is we have
developed machine learning algorithms that are able to analyze human behavior
in terms of analyzing medical images and then simulate that behavior in
response to new images that the algorithm has not seen before.
DR. JASON BIRNHOLZ:
My thanks to Dr. Roee Lazebnik, who has been our guest. We
have been discussing workflow efficiencies and ultrasound. Roee, thank you
very much for sharing your work and workflow concepts with us.
DR. ROEE LAZEBNIK:
It was my pleasure too.
DR. JASON BIRNHOLZ:
I am Dr. Jason Birnholz. Please visit our website at www.reachmd.com, which features our entire
library through on-demand podcasts.
You have been listening to advances in medical imaging
sponsored by Siemens Ultrasound. For more details on this week’s show or to
download the segment, visit us at www.reachmd.com.
Thank you for listening.
Siemens Acuson Ultrasound Systems feature intelligent
knowledge-based algorithms that automate and screen line a broad range of
clinical tasks. It’s a whole new level of accuracy and efficiency. Siemens
Ultrasound knowledge-based workflow technology enables you to achieve
consistent and reproducible measurements with high speed and stability across a
range of applications and users. Knowledge based workflow applications on
Siemens Acuson Ultrasound Systems are different than conventional image
processing software. Siemens leverages an experienced based and contact
specific artificial intelligence approach to automate many routine clinical
imaging tasks. Using learnt pattern recognition from an expert database of
thousands of real critical cases, knowledge based workflow applications
recognize anatomical patterns and landmarks and can automatically perform
measurement. As a result, you spend less time interacting with an image volume
and more time with the clinical information it contained. You experience less
variability and you have access to more accurate quantitative image-based
measurement from prevention and early detection to diagnosis and therapy.
Siemens Healthcare is the only company that delivers a full spectrum of
healthcare solution. Siemens, answers for life. For more information, visit www.siemens.com/ultrasound.
Facebook Comments