14

My question is how to measure very small movements of a needle and syringe when injecting. When doctors inject local anaesthetic, they always aspirate (suck back) first to make sure they are not in a blood vessel. My contention, particularly if the aspiration is done single-handed, is that the change in direction of forces on the needle/syringe combination whilst aspirating causes significant movement of the end of the needle - probably several mm - which negates the purpose of aspiration in the first place.

I want to do an in vitro study, in which I have a needle and syringe combination and inject into a piece of meat or similar - and then get volunteers to aspirate / inject under 3 circumstances:

  1. stabilise with other hand and inject directly
  2. stabilise with other hand, aspirate and then inject
  3. aspirate with one hand and then inject

I have hit a block in terms of finding a method of measuring these movements of the needle tip down to maybe 0.1mm. I thought that an accelerometer might be the way but have not found anything small enough to be mounted on the needle tip.

The only other way I thought of doing it was to use a camera mounted side on to the tip of the needle which would be protruding through some sort of artificial 'skin' and then have a graticule calibrated to measure the distance moved.

Mark Stoneham
  • 161
  • 3
  • 5
    Can please you edit your question so that it only contains your specific electronics problem? The question is very broad and very little of this seems to have anything to do with electronics - how to best mount an accelerometer on a moving object is a mechanics question. – Lundin Sep 12 '16 at 09:33
  • 5
    IMO there is a good electronics under there though - accurate sensing of the 3D movement of a small object. I think choosing sensors and designing a measurement is reasonably on-topic. – Jack B Sep 12 '16 at 09:44
  • 1
    Looking at your post, you clearly didn't intend a wall of text, but your attempts at formatting were not recognised by the site software. Edited so your formatting does what I think you intended (2 spaces at end of line in lists for example) – Neil_UK Sep 12 '16 at 10:00
  • 1
    Do you want absolute position of the needle, or position relative to the surface of the skin? Do you want all 3 axes, or would some kind of rangefinder on the axis of the syringe be useful? – pjc50 Sep 12 '16 at 10:46
  • 2
    I'm confused. Are you trying to sense the 3D position of the tip of the needle, or are you really only trying to measure how far the needle is inserted into the patient or slab of meat? – Olin Lathrop Sep 12 '16 at 11:06
  • FWIW, if you just want to locate the needle tip linearly (depth), my idea would be to attach an electrode to one end of "meat", the other to the needle, and measure the resistance across with a precision (uOhm range probably) multimeter. Assume the "meat" is a linear conductor, so the Rtot = Rneedle (which is about constant and probably can be brought near 0 when using e.g. a custom silver/gold needle) + meat res. per length unit * "meat" length. This boils down the problem to a linear regression one. I can't guarantee the results (IANAB), but take a try and check if the idea works for you. –  Sep 12 '16 at 12:51

7 Answers7

16

An accelerometer is definitely out in terms of noise.

A mechanical arm system, while potentially accurate enough, may well influence the injection scenario enough to render your results meaningless. I suspect a student struggling to control the position of a small syringe would be distracted by a large measurement arm, no matter how well balanced and low friction.

The only real \$^*\$ options you have are optical.

It should be possible to mark the syringe at both ends of the barrel with fiducial markers. The resolution you can achieve is limited by the optics for pointing multiple cameras at the target. If the test site is small and the location well defined, then you can use zoom optics to make the image fill a significant amount of the frame. HD cameras, and sub-pixel location of the fiducials via something like OpenCV ought to make your target resolution achievable.

\$^*\$ real => low cost, keeps the imaging volume clear, and it's apparent how you'd get the resolution. There are plenty of other modalities, for instance MRI, PET tomography, ultrasound, magnetic tomography, Xray CT, resistive tomography - which need variously calibration, development, expensive equipment etc.

Neil_UK
  • 158,152
  • 3
  • 173
  • 387
  • 2
    To add to this, if you use ballistic gel or some similar medium instead of meat, then it's transparent. With the needle coloured black (Sharpie pen?), it should show up pretty well on camera. That way you don't even have to calculate the tip position from the syringe barrel - you can see it directly in the image. – Graham Sep 12 '16 at 13:55
  • 2
    @Graham I would be afraid of giving the injector unrealistic feedback by making the needle tip visible to them. – Wayne Conrad Sep 12 '16 at 19:50
  • Ballistic gel on a glass plate, imaged from below, with a top layer loaded with carbon black so it's opaque? (Also @Graham) – Chris H Sep 13 '16 at 08:02
  • But if you go IR and the layer isn't too thick you'll get appreciable transmission in flesh and still be able to pick it up on a silicon CCD. – Chris H Sep 13 '16 at 08:05
  • @Neil_UK Your solution seems to be viable with reasonable effort and well explained. But claiming all other options beeing unviable or unthrifty just by throwing in some buzzwords sounds a little bit overconfident, doesn't it? Don't get me wrong. It may be as you write and all other methods are way worse, but to write "real *" the way you do, you should offer some proof or explanation. – Ariser Sep 13 '16 at 09:34
  • 1
    @WayneConrad If you're simulating an injection, then you need simulated skin on the top. Having stuck needles in someone, I know there's a definite difference in going through skin and going through the underlying flesh. A simple sheet of thin rubber might be enough - and that'll be enough to stop the injector seeing the tip. Injection simulation equipment already exists (a quick Google found http://www.wallcur.com/Products/Practi-Injecta-Pad-Injection-Simulator__501PD.aspx) so perhaps use that as a starting point. – Graham Sep 13 '16 at 09:43
6

I would err away from an accelerometer. To get displacement from an accelerometer means intergrating twice - once to get the velocity, and again to get the position. This means errors tend to accumulate. Also, the accelerometer would need to be attached to the needle, and it would probably be better to use a needle and syringe which looks and feels as normal as possible to the volunteer.

The idea with the camera sounds better to me. You probably want to measure all the axes of motion of the needle and syringe - so all three movement directions and all three rotation directions. You should be able to do that fairly easily with two cameras, one looking across the surface of the "skin" the other looking down from above. Put the cameras well back from the subject and use a long lens, that will reduce perspective effects. If you put a few brightly coloured dots on a black-painted syringe, it'll be fairly easy to track them in a video using something like ImageJ and a tracking plugin. You can then use the movement of those points to reconstruct the movement of the whole syringe.

Jack B
  • 12,167
  • 1
  • 27
  • 46
  • 2
    You probably want a similar dot pattern grid on the surface of the meat, perhaps on a thin membrane like Tegaderm, to calibrate its movement and deformation. –  Sep 12 '16 at 10:08
5

Note that setting up this study will be a big deal. Getting the instrumentation right will be a study in and of itself. I suggest poring over Medline or Google Scholar for a few days to see if anyone has done anything similar. Look at papers you've read that lead you to this area of research for guidance.

Personally, I'd be looking toward ultrasound imaging to tell you what you need to know. In fact, I suspect this is how anesthesiologist typically guide the needles, and if the tip were moving too much during aspiration, they'd probably already know it.

I don't like "google is your friend" - type messages, but I'll add that searching for "needle motion measurement" in google scholar yields tons of hits, and the first hit for me indeed points to ultrasound: http://scitation.aip.org/content/aapm/journal/medphys/33/8/10.1118/1.2218061

Scott Seidman
  • 29,274
  • 4
  • 44
  • 109
  • They used an Acuson 128 XP ultrasound (presumably with the highest available MHz transducer) and a [6 DOF force-torque sensor](http://www.ati-ia.com/company/NewsArticle.aspx?id=310182740) and a [magnetic position sensor](http://www.mindflux.com.au/products/ascension/minibird.pdf). The latter seems a bit marginal for the OP (0.5mm RMS resolution/1.8mm RMS static accuracy). – Spehro Pefhany Sep 12 '16 at 15:12
  • 1
    @SpehroPefhany -- I wasn't trying to point to *the* answer with that link, but to clue the OP in that there *is a literature*. That paper will have cited other papers, other papers will have cited it, and the OP needs to use tools to do both sorts of searches. That's how research is done. – Scott Seidman Sep 12 '16 at 16:06
  • 1
    The particular paper you pointed to was unusually helpful. Usually I shell out $35 and end up with disappointment. – Spehro Pefhany Sep 12 '16 at 16:07
4

I object that the only means were optical or fix mounting to a measurement system. I propose a different approach:

Resistive sensing

Alongside the blood vessel attach a strip with an array of electrodes. On the syringe a small voltage is applied. After calibration and developing a good algorithm I think it is possible to determine the position of the tip of the needle accurately by analysing the change of current/voltage at the different electrodes. To keep physiological effects low you might use RF voltages above 300 kHz. This also allows the evaluation of the phase increasing reliability of this method.

drawback: The area where the test should take place may be obscured by the strip.

inductive sensing

Imaging a number of inductive sensors placed in vincinity of a blood vessel. They should be able to deliver 3D data of the needle.

Both approaches involve ample development work. Though the outcome may be a system capable of replacing the aspiration process at all.

Community
  • 1
Ariser
  • 3,846
  • 3
  • 23
  • 43
3

There are many potential methods.
Refining the problem will improve the chances of find a good system.

If needle movement is relative to target (meat, person etc) so that net length of needle inside target varies then measurement of resistance (DC, AC, ...) of needle to target path may allow motion to be detected. The absolute value will almost certainly not be repeatable accurately between "runs" but delta-motion should be observable for very small movements.

If passage of needle relative to target surface occurs during draw back then a sensor that measures absolute needle position relative to target plane at entry point could be achieved by various means. One of several is to have a "disc on the needle near the surface and measure capacitance change - small in absolute terms but doable.

If the injection simulator may be somewhat unrealistic in appearance while accurately modelling the key required actions you could telemeter motion woth hall cells or GMR sensors and magnets, LVDT sensors , mechanical couplings, ... .

ie it seems mainly a matter of really understanding the fine detail of exactly what you want to measure and then coupling one of a number of sensors to it.

Russell McMahon
  • 147,325
  • 18
  • 210
  • 386
2

I don't think there is an electrical design question in your post but I think you're looking for a 6-axis coordinate measurement machine. This is like the familiar 6-axis robots used in car-manufacturing but has no motors - only encoders to read the position.

enter image description here

Figure 1. A Mitutoyo Spin-arm Apex Series 6-axis coordinate measurement system.

I didn't read the doc but you would need something with loose joints so that the measurement device doesn't stabilise the needle or add any inertia to the operation.

Ariser
  • 3,846
  • 3
  • 23
  • 43
Transistor
  • 168,990
  • 12
  • 186
  • 385
0

Sensors to consider: LVDT, strain gage, piezo. Sensing on all of these would have to be done at the needle-mount end, not at the tip. I'm assuming that you may modify your "test-needle" with this instrumentation...there should be enough room inside the needle body (hand-held part) to accommodate sensors like this. Mechanical mounting and coupling of these sensors are not topics appropriate for this site.
Of these, LVDT sensors are very robust, and can be impervious to a liquid environment. And many have excellent sensitivity. One-of-many example: http://www.disensors.com/downloads/products/MHR%20Miniature%20LVDT_521.pdf

glen_geek
  • 23,591
  • 1
  • 22
  • 50