Skip to main content
. Author manuscript; available in PMC: 2014 Feb 27.
Published in final edited form as: JAMA Intern Med. 2013 Oct 28;173(19):1788–1796. doi: 10.1001/jamainternmed.2013.9245

Table 4.

Accuracy of individual ultrasound characteristics.

Characteristic Level Cancers
Controls
Likelihood Ratio
Odds Ratio (95% CI)
Sensitivitya
No. (%)(n = 102)
False Positive Ratea
No. (%)(n = 515)
Characteristic Present
LR (95% CI)
Characteristic Absent
LR (95% CI)
Microcalcifications 39 (.38) 28 (.05) 7.0 (6.0–8.2) 0.65 (0.56–0.76) 11.6 (6.5–20)b

Coarse calcifications 13 (.13) 34 (.07) 1.9 (1.6–2.3) 0.93 (0.78–1.1) 2.1 (1.1–4.1)c

Rim Calcifications 5 (.05) 23 (.04) 1.1 (0.91–1.3) 1.0 (0.82–1.2) 1.0 (0.41–2.6)

Comet Tail Artifact 17 (.17) 77 (.15) 1.1 (0.93–1.3) 0.98 (0.82–1.2) 1.1 (0.60–2.0)

Composition
 Entirely Solid 68 (.67) 220 (.43) 1.6 (1.4–1.8) 0.58 (0.51–0.66) 2.2 (1.4–3.5)b

 Entirely Cystice 0 (.00) 37 (.07) 0.034 (0.028–0.041) 1.1 (0.88–1.3) e

 Mixed Solid and Cystic 34 (.33) 248 (.48) 0.69 (0.58–0.82) 1.3 (1.1–1.5) 1

Echogenicity
 Hypoechoic to strap 16 (.16) 33 (.06) 2.4 (2.0–2.9) 0.89 (0.75–1.1) 2.9 (1.4–6.0)d

 Isoechoic/hyper to strap 51 (.50) 198 (.38) 1.3 (1.1–1.5) 0.81 (0.70–0.94) 1.8 (1.1–2.9)c

 Isoechoic to thyroid 30 (.29) 209 (.41) 0.72 (0.61–0.86) 1.2 (1.0–1.4) 1

 Hyperechoic to thyroid 4 (.04) 38 (.07) 0.53 (0.44–0.64) 1.0 (0.86–1.3) 0.75 (0.25–2.2)

Central vascular flow 40 (.39) 136 (.26) 1.5 (1.3–1.7) 0.83 (0.71–0.97) 1.6 (1.0–2.6)c

Peripheralvascular flow 49 (.48) 201 (.39) 1.2 (1.1–1.4) 0.85 (0.73–0.99) 1.3 (0.82–2.1)

Halo 28 (.27) 116 (.23) 1.2 (1.0–1.4) 0.94 (0.79–1.1) 2.0 (1.3–3.1)

Margins (ill ldefined/lobulated) 61 (.60) 212 (.41) 1.5 (1.3–1.7) 0.68 (0.60–0.78) 2.0 (1.3–3.1)c

Shape (Taller than wide) 18 (.18) 42 (.08) 2.2 (1.8–2.6) 0.89 (0.75–1.1) 2.3 (1.2–4.3)c

Size
 <=1cm 30 (.29) 248 (.48) 0.61 (0.51–0.73) 1.4 1

 1.01–2cm 38 (.37) 169 (.33) 1.1 (0.97–1.3) 0.94 1.9 (1.2–3.1)d

 >2cm 34 (.33) 97 (.19) 1.8 (1.5–2.1) 0.82 3.1 (1.8–5.2)b
a

Each characteristic was considered individually and was not adjusted for the presence of the other findings. The odds ratios were adjusted for multiple nodules within subjects, while the accuracy statistics were not adjusted for multiple nodules within subjects. The numbers may not sum as nodules in which the ultrasound characteristic was missing are included in the calculations of sensitivity and specificity but are not shown in the table.

b

p < .001

c

p < .05

d

p < .01

e

Entirely cystic includes cystic honeycomb. To calculate the likelihood ratio, 0.25 was added to zero cells. Because entirely cystic nodules had no cancer cases, an odds ratio could not be calculated.