-
Measuring to the sub-millimeter
Newest Lung-RADS criteria recommends measuring nodules to the tenth of a millimeter. I think this is absurd. Take your fingers, and look at what a millimeter looks like. Then try to fractionate that into tenths. Now think about slice selection on axial imaging and reproducibility between those doing the measuring.
It can’t be possible to be precise and reproducible to measure pulmonary nodules down to tenths of a millimeter, can it?
(As a side note, in my experience, ultrasound measurements by the technologists are ridiculously inaccurate when measuring things that are on the order of a couple millimeters. I don’t know how many times the tech has measured the CBD at 8mm, but when I use my calipers it is 5mm. Its not typically a fault of their own, but a fault of the imprecise measuring tools on their machine).
end rant. -
19 Replies
-
I don’t do body imaging and have no idea about Lung-RADS, but I have never reported any case to a tenth of a mm in my career. No chance anyone can be that precise with standard imaging.
At most, I would say “4-5 mm” but I would never say 4.3 mm. It’s a made up number and it’s not good for patients to make things up.
-
You mean to say that saying a subpleural nodule that ‘grew’ from 2.2mm to 2.4 mm over one year isn’t important? /s
-
Quote from ultimaterads
You mean to say that saying a subpleural nodule that ‘grew’ from 2.2mm to 2.4 mm over one year isn’t important? /s
If you think it’s definitely grown then yes you should say that.
But if you think you are able to measure things accurately to a 10th of a mm, I would disagree. If you measure the same nodule on the same study on the same day you could easily get 2.2 mm one time and 2.4 mm another time.
I cringe when I follow something that’s identical but supposedly grew a tenth of a mm over a year or two. It’s nonsense.
-
And this is why people hate radiologists.
-
Quote from Radsoxfan
Quote from ultimaterads
You mean to say that saying a subpleural nodule that ‘grew’ from 2.2mm to 2.4 mm over one year isn’t important? /s
If you think it’s definitely grown then yes you should say that.
But if you think you are able to measure things accurately to a 10th of a mm, I would disagree. If you measure the same nodule on the same study on the same day you could easily get 2.2 mm one time and 2.4 mm another time.
I cringe when I follow something that’s identical but supposedly grew a tenth of a mm over a year or two. It’s nonsense.
Yeah its dumb to think one can be consistently accurate to the tenth. Hence the sarcasm tag.
-
Lung rads is becoming like birads. Use our terminology and adhere to our arbitrary decisions or else!
-
-
-
-
Agree. Absurd. This is not comparable on subsequent imaging due to differences in imaging technique making its original purpose obsolete.
-
Used to have a nodule tool on the AW workstation. It did some histogram thing to set the calipers and measure nodule volumes. If you followed patients on the same scanner with the same protocol, the measurements meant something.
-
This is why the word “similar” is my favorite word to describe something that is grossly similar given differences in technique and measurement.
-
I use no significant interval change
-
One could argue measuring to the mm would make you more precise even with no effect on accuracy (using this in terms of the statistical meaning). I think most radiologists and many pulmonologists are aware measurements can differ among rads by 1-2mm. So say there is a 4.5 mm nodule, the true range of measurements for that nodule is 2.5-6.5, because of how wide a range of measurements can be produced. But, if you round it and report it as 5mm, you are suggesting the range is 3-7 mm, that the point estimate is different than what you are actually measuring, due to some inaccuracy introduced by rounding.
Just playing devils advocate here
-
Ridiculous. No way measurements to a tenth of a millimeter are accurate. Was hoping we were headed in the right direction with RECIST and revised fleishner, but clearly a few academics just needed some more publications. Pathetic disconnect between academia and the real world.
-
Agree with general comments.
With the range of error, you can always remeasure to get the number you want…
Similar is a favorite word.
Can eyeball 2-3 mm, 3-4 mm etc pretty accurately.
I’ll sometimes use measurements as justification for followup on minimally enlarging lesions when I’m just worried about it and want to keep on top of it.
-
absurd. On a CT with slice thickness of a 1-2 mm it is crazy and inaccurate to report to the tenth.
It reminds me of signing out with a young OCD/paranoid neurorad attending in my residency. I had to completely change my reports if I found out he was the sign out. He was the type to report findings to the tenth of a mm. Calling possible SAH on every CT head trauma cases (especially the ground level fall drunks) to buy them a night in the ICU. He caused way more harm than good with all his overcalling.-
The responders so far are all in agreement.
But the ones who will write the textbooks say otherwise.
Anyone want to take up the other side of the argument?
-
I interpreted that as more of 5-6mm / 6-7mm than actual tenths. Because obviously, you can’t get that level of accuracy.
-
Same applies to renal sizes given to the mm on ultrasound.
-
I thought I presented the counter argument (not that I completely agree with it) but will restate:
The issues of measuring the sub-mm and the variability in measurements are separate issues. One (sub-mm measurements) addresses accuracy the other (variability) addresses precision. One could argue measuring to the .5 mm would make you more precise even with no effect on accuracy. Say measurements can vary among rads by 1 mm. So say there is a 4.5 mm nodule, the true range of measurements for that nodule is 3.5-5.5. But, if you round it and report it as 5mm, you are suggesting the range is 4-6 mm, due to some inaccuracy introduced by rounding. Well the second range encompasses the 6mm breakpoint and could inaccurately result in a follow up study by a pulmonologist or someone who is thinking in terms of the variability of measurements.
One way we could think about this is in terms of reporting ranges instead of point estimates. So 2-3 mm nodule or 1.5-2.5 mm nodule. If I am measuring the nodule at 2mm I prefer the latter range -
[blockquote]”Given that diameter measurements vary by 1.73 mm across observers for nodules smaller than 2 cm ([link=https://pubs.rsna.org/doi/full/10.1148/radiol.2017162894#r26]26[/link]), it appears reasonable to report growth when a change in measured diameter of at least 2 mm is detected (actually at least 1.5 mm due to rounding). Use of this 2-mm threshold would reduce the likelihood of an incorrect diagnosis of growth when the apparent difference in size is in fact within an expected range of uncertainty owing to observer-related imprecision.”
[/blockquote] [size=”3″][link=https://pubs.rsna.org/doi/full/10.1148/radiol.2017162894]Recommendations for Measuring Pulmonary Nodules at CT: A Statement from the Fleischner Society[/link] [/size]
[size=”3″]
[/size]
-
-
-
-
-
-
-
-
-
