Is Notified Body (NB) audit required for Class I imaging device ?

[pkost]

There was a recent case which progressed through the european courts that confirmed that the manufacturers intended use is the intended use.

Therefore if the manufacturer does not claim measuring, it shoudl not be assessed as such - because it is not!

Guidance on the application of "measuring" functions (MEDDEV 2.1/5) gives 3 criteria for a product to be a measuring device....

1. The device is intended by the manufacturer to measure
2. The result of the measurement is displayed in legal units, or other acceptable units, or is compared to at least one point of reference indicated in legal units etc etc
3. the intended purpose implies accuracy, claimed explicitly or implicity ... ...


If a imaging device by a manufacturer is not intended to meausure and there is no scale to make a measurement, it is not appropriate to declare as a measuring device.

 

[sagai]

I must admit it is a surprise for me that any imaging device could be a Class I kind of thing and I deeply agree with Miregmgr.

Can we ask the OP what does this device actually do for imaging though?

Cheers!

 

[sagai]

Measuring can be a simple yes/no regardless there is no explicit measuring function for the imaging application.
One obvious case to be mentioned, aneurysm is either on the image or not on the image, simple is that without any explicit measuring function it drives the physician to cut the patient. Moreover measuring can be relative measuring based on reference body particle on the same image.
Cheers!

 

[MIREGMGR]

For discussion:

Those with medical imaging technical knowledge will be aware that anatomical-feature image elements may be displayed at their actual size, or larger or smaller relative to the anatomy around them, according to the incident beam source size, technical characteristics of the detector, and either intentional response nonlinearity of a film/phosphor or fluorophosphor display to provide edge sharpening, or quantitative processing of the image data from an electronic detector prior to display so as to similarly provide edge sharpening.

Edge sharpening processes have long been understood to be medically useful, in greatly easing the image reader's task of discerning the medical significance of minor density gradations...but edge sharpening processes always have the potential to cause anatomical features to be displayed not at true size.

The most important features to be diagnostically evaluated in mammographic imaging...cysts, duct conditions, self contained and invasive carcinomas...present as features on images. A key part of diagnosis is size scoring. The image process itself doesn't measure sizes, but it must present images that display features at close to true size, notwithstanding whatever edge sharpening technology is utilized, so that the diagnostician can correctly evaluate, measure and thus score the observed features.

My understanding is that when FDA evaluates a mammographic imaging system 510(k) for acceptable performance, they give heavy consideration to proof that the resulting images support correct measurement. I don't know if a parallel approach is utilized in the EU, but I would think that would be the case.

It's true that the diagnostician with his or her transparent-Mylar measuring scale is performing the direct measurement, but that measurement process will be useless if the imaging system displays features as larger or smaller than their actual size, because the clinician has no way to separately obtain true-size information. Thus the patient may be mistreated and harmed.

Similar considerations apply, at least with FDA, to CT systems that are used, not to create visual diagnostic displays, but rather to create 3D measured-dimensional data of a tumor location relative to marker features on a patient...that data then later being used to directly control the beam scanning process of an X-ray, E-beam, proton or neutron scanner that destructively irradiates the tumor while sparing the surrounding anatomy.

If the CT scanning and data creation process gets the measurements wrong, the tumor will not be correctly treated, and the patient will be harmed by destruction of healthy anatomy.

My understanding is that in these and other imaging contexts, FDA regards imaging as inherently a measurement process, irrespective of whether there is a calibrated scale in the particular device.

 

[Ronen E]

Measuring can be a simple yes/no regardless there is no explicit measuring function for the imaging application.
One obvious case to be mentioned, aneurysm is either on the image or not on the image, simple is that without any explicit measuring function it drives the physician to cut the patient. Moreover measuring can be relative measuring based on reference body particle on the same image.
Cheers!

I think you're confusing measuring with diagnosing. As far as I understand measuring must yield an absolute numeric result (excluding Boolean 0 & 1).

 

[Ronen E]

...

It's true that the diagnostician with his or her transparent-Mylar measuring scale is performing the direct measurement, but that measurement process will be useless if the imaging system displays features as larger or smaller than their actual size, because the clinician has no way to separately obtain true-size information. Thus the patient may be mistreated and harmed.

Similar considerations apply, at least with FDA, to CT systems that are used, not to create visual diagnostic displays, but rather to create 3D measured-dimensional data of a tumor location relative to marker features on a patient...that data then later being used to directly control the beam scanning process of an X-ray, E-beam, proton or neutron scanner that destructively irradiates the tumor while sparing the surrounding anatomy.

...

My understanding is that in these and other imaging contexts, FDA regards imaging as inherently a measurement process, irrespective of whether there is a calibrated scale in the particular device.

I can see the importance, but aren't these aspects already covered by the requirement that the device will perform it's intended use (e.g. obtaining real-size image) safely and effectively, e.g. within X% of real-size in the examples above?

I'm not going to argue whether or not this is the actual FDA's practice because I have no such specific experience (and to me, this whole discussion is less relevant to the USA (1) because the OP enquired about the EC and NB involvement, and (2) the USA system doesn't have the equivalent of "1m" class); I only express my opinion that this is a too-conservative approach and actually a redundant measure (pun not intended).

Cheers,
Ronen.

 

[sagai]

Would you consider a digital thermometer saying "yes" for fever (above 37 Celsius) and "no" for anything under 37 Celsius as a non measuring device?
Anyway, this is off topic already because Class I device can have a measuring function, please find:
http://ec.europa.eu/health/medical-devices/files/meddev/2_1_6_ol_en.pdf

Cheers!

 

[Ronen E]

Would you consider a digital thermometer saying "yes" for fever (above 37 Celsius) and "no" for anything under 37 Celsius as a non measuring device?
Anyway, this is off topic already because Class I device can have a measuring function, please find:
http://ec.europa.eu/health/medical-devices/files/meddev/2_1_6_ol_en.pdf

Cheers!

That thermometer actually conveys an absolute numerical measurement result, even though it does not display a numerical figure. Definitely that result is present in the internal workings of that thermometer. I bet that the resulting accuracy and precision of those workings will be scrutinized up close in any regulatory clearance process. It's quite different to the general notion of "providing a true visual represantation of an internal body structure / organ". How would you pin-point the accuracy and precision of that?

Of course there are class 1 devices with a measuring function, I don't think anyone in this thread suggested otherwise. The question is whether the NB has the authority (not just the technical ability) to force an "intended" measuring function when it's actually NOT intended by the manufacturer. It's mostly interesting in the case of class 1 devices, becuase in that case it makes a difference (requiring / not requiring NB involvement). With >1 class devices it makes a lesser difference, at least in the context of the OP's question.

Cheers,
Ronen.

 

[sagai]

So what is a digital image?
It shows the detected electromagnetic ray intensity as numerical values based on algorithms built in the software and displays this intensity as an image.
The point is regardless we see "only" an image, it is a set of numerical values anyway.
With other words, the equivalent output of the image also could be a matrix of numbers that represents converted numerical values of this projected intensity, it is not a question whether or not it measures something (because of course it does, the intensity itself) rather a question of how does it to be displayed for human.

However I am still curious to know more about this imaging application that is deemed to be class I.

Cheers!

 

[Ronen E]

So what is a digital image?
It shows the detected electromagnetic ray intensity as numerical values based on algorithms built in the software and displays this intensity as an image.
The point is regardless we see "only" an image, it is a set of numerical values anyway.
With other words, the equivalent output of the image also could be a matrix of numbers that represents converted numerical values of this projected intensity, it is not a question whether or not it measures something (because of course it does, the intensity itself) rather a question of how does it to be displayed for human.

This is what I call wordsmithing. I give up.