You can find articles in the medical literature that claim both?
For the most part and for most of the time that CTA (computed axial tomographic angiography, in this context pulmonary artery CTA) ( AKA pulmonary CT Angiography) has been used the consensus number for the false positive rate is about 5%.
Two more recent articles provide data that strongly suggest that the real false positive rate may be considerably higher than that.
Hutchinson et al published an article in the American Journal of Radiology in August 2015 entitled Overdiagnosis of Pulmonary Embolism by Pulmonary CT Angiography" (august 2015,vol.205, number 2, 271-277)
In the study all pulmonary CTAs read-as- positive for pulmonary emboli (PE) done in a referral university hospital over a 12 month period were reviewed . There were 174 cases originally reported as positive for PE. The images were reviewed by a panel of three chest radiologists with over ten yeas of experience. I assume by each.
Those cases in which all three of the radiologists agree that the studies were negative for PE were considered false positive. 45/174 were false positive or 25.9 %,
Cases which were considered false positive tended to be in the segmental or subsegmental vessels .
According to the three judge panel the reasons for false positivity were breath holding artifact and beam hardening artifact.
A previous article published in CHEST in 2009 ( Miller wt et al , Frequency and Causes of false-positive CTPA exams in community Hospitals, Chest 2009,,136 4,Meeting Abstract 145) reported a value of 11% for false positive tests.This study , although it involved large numbers of cases, was much less methodologically sound and , perhaps for that reason attracted not much attention. 608 cases originally read as positive for PE were reviewed by a single radiologist who found that 11% were "either false positive or probably false positive". Most ( 82%) were judged by the reviewing radiologist to be due to technical factors, which included" pulsation artifact,streak artifact,volume averaging, and quantum mottle". 66% of the false positive cases were at the subsegmental level.This article seemed mainly to be a he said she said type exercise with one radiologist compared with a number of other radiologist's film readings.So how do you know who is right?
It seems well known that technical factors are the major reason for false positive readings and analyzing findings at the subsegmental level. is particularly difficult.
A widely read and quoted article from the NEJM by Stein in 2006 quote the value of 5% false positive rate and I suspect the 5% value, more or less, is what most physicians accept as correct.
These data are derived from the famous PIOPED 11 study. The CTAs were read by two radiologists and the process to determine positive and negative was rigorous and detailed and can be found in the method section of the article. However it is not the case that the CTA was directly compared to the then gold standard of pulmonary angiogram,which ,of course, was also not done in the Hutchinson paper. Stein's paper reported a specificity of 95% .
It is easy to discount the Miller paper but the Hutchinson paper is disconcerting particularly in light of my recent health events.I was diagnosed by CTA as having several PEs.I had the study re-read by a long time friend who is a radiologist who said he could detect no emboli.To "break the: tie I had my treating physician to review the study with a radiologist of his choice who concluded that there were several artifacts present that could be mistaken for emboli but that there were "one or two" he said likely were clots. My doctor and I agreed to continue the apixaban for the agreed upon 3 months ( this was a provoked embolic situation).More on that "teaching moment " later.
Addendum: 1/19/16 Overdiagnosis and false positive are related but distinctly different concepts. Overdiagnois refers to finding something that is really there but does not really need treatment and false positive refers to thinking you are finding something but it is not really there.Subsegmental emboli in CTPA may exemplify both concepts, particularly i n the case of the reading of a single subsegmental embolus. See Dr. Samuel Goldhaber's comments from Medscape found here.