I assume "per CT scan" means one patient's CT scan at a point of time. Each CT scan contains approximately 200 CT photos with three directions as axial, coronal, and sagittal. For example: "They achieved 100% sensitivity for lung nodules larger than 3 mm, and 70% sensitivity for lung nodules less than or equal to 3 mm at 15 FPs per scan". I am not sure what does it mean by 15 false positives per scan. Can anyone explain?
Original journal: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2800987/
When a radiologist says there are n false positives per CT scan, what does it mean?
2019-11-02 2:21 am
回答 (4)
2019-11-02 4:50 am
A positive event is detection of an object that is counted as a nodule. Each scan can have none or may nodules. Fifteen false positives per scan means that on average scan, 15 objects were counted as a nodule but were not actual nodules.
2019-11-03 12:19 am
What this means is that what shows up as a bright spot in the scan may or may not be a nodule, typically we have to balance between false positives and false negatives in statistics.
On the pure statistical standard, as you aim to decrease the number of false positives, you likewise increase the number of false negatives... decreasing the number of false negatives increases the number of false positives. This holds true unless you have a change in methodology (meaning you change the way the data is collected or the quality of the data collected).
For each "photo" taken in a CT scan, you really only know the amount of energy that got absorbed and the direction it was absorbed in (since the emitter is uni-directional and only turns on at a given part of the scan). The scan as a whole is what has the false positives, the individual slices aren't very useful, but requires a large amount of calculations to create something that is usable to doctors.
Hence what your article seems to be describing is a change in methodology that lowers the number of false positives for nodules under 3mm. A "bright spot" normally is what we're looking for in the scan, so the change in methodology makes it easier to tell that the spot isn't due to other issues like scarring, patient movement, or other such issues... but instead because it is an actual cancer nodule.
On the pure statistical standard, as you aim to decrease the number of false positives, you likewise increase the number of false negatives... decreasing the number of false negatives increases the number of false positives. This holds true unless you have a change in methodology (meaning you change the way the data is collected or the quality of the data collected).
For each "photo" taken in a CT scan, you really only know the amount of energy that got absorbed and the direction it was absorbed in (since the emitter is uni-directional and only turns on at a given part of the scan). The scan as a whole is what has the false positives, the individual slices aren't very useful, but requires a large amount of calculations to create something that is usable to doctors.
Hence what your article seems to be describing is a change in methodology that lowers the number of false positives for nodules under 3mm. A "bright spot" normally is what we're looking for in the scan, so the change in methodology makes it easier to tell that the spot isn't due to other issues like scarring, patient movement, or other such issues... but instead because it is an actual cancer nodule.
2019-11-02 3:11 am
I have cancer of the anus. I waited too long to see a doctor so it spread to my balls
2019-11-02 3:09 am
It means don't be an idiot and wait to talk over your results with your doctor.
收錄日期: 2021-05-03 21:48:51
原文連結 [永久失效]:
https://hk.answers.yahoo.com/question/index?qid=20191101182145AAmhqWR