Clinical Question

What is the likelihood and what are the outcomes of incidental findings on imaging tests?

Bottom Line

The risks of imaging, in addition to radiation exposure, include the identification of incidentalomas, which can lead to patient anxiety, further testing, and overtreatment. There is little research to guide what to do when they pop up on an imaging report (as the famous dodge “clinical correlation needed”). Computed tomography (CT) of the chest (45%), CT colonoscopy (38%), and cardiac magnetic resonance imaging (MRI) (34%) commonly produce incidental findings. The rate of malignancy in incidentalomas was high in breast (42%) and ovary (28%) findings; intermediate in prostatic and colonic (10% to 20%) findings; and low in brain, parotid, and adrenal gland (less than 5%) findings. Although everyone has a story of the lifesaving results of such serendipity, we do not often consider the patients subjected to unneeded testing and treatment, the so-called victims of modern imaging technology—you can figure out the acronym (BMJ. 2003;326:1273). (Level of Evidence = 2a)

Synopsis

These authors searched two databases and reference lists of included papers to identify 20 systematic reviews of observational studies that gave a prevalence of incidental abnormalities (incidentalomas) in patients already being imaged for cancer. Incidentalomas were defined differently across the systematic reviews. CT of the chest resulted in incidentalomas reported in 45% of patients (95% confidence interval [CI], 36% to 55%). The relatively new CT colonoscopy resulted in incidental findings in 38% of patients (21% to 57%). MRI also reported incidental findings when imaging the spine (22%) and brain (22%). Whole body positron emission tomography (PET) and PET/CT had rates of 2% (95% CI, 1% to 4%). No studies have determined the prevalence of incidentalomas identified via radiography or ultrasonography. Malignancy of incidentalomas were highest with breast findings (42%; 95% CI, 31% to 54%). Renal, thyroid, and ovarian findings were malignant approximately 25% of the time. Extracolonic, prostatic, and colonic incidentalomas were malignant 10% to 20% of the time. Rates of incidentalomas varied substantially among the meta-analyses.

Study design: Systematic review

Funding source: Self-funded or unfunded

Setting: Various (meta-analysis)

Reference: O'SullivanJWMuntingaTGriggSIoannidisJPAPrevalence and outcomes of incidental imaging findings: umbrella review. BMJ2018;361:k2387.