How to Dictate a Thyroid Ultrasound: TI-RADS Without the Cheat Sheet

The thyroid ultrasound is where structured reporting stops being optional. Every nodule you describe will be scored — by you, by the next radiologist, or by an endocrinologist reading your prose with the ACR TI-RADS table open in another window. A report that describes nodules in TI-RADS vocabulary scores itself. A report that says "heterogeneous nodule, probably benign" forces someone else to guess which boxes you meant to check, and the guess changes whether the patient gets a needle.

The gland first

Before any nodule, the gland:

• Each lobe in three dimensions — length, width (transverse), and depth (AP) — in centimeters. Normal adult lobe length is roughly 4–6 cm. Two dimensions is a habit attendings correct; the third is what makes a follow-up comparable.
• Isthmus thickness, normally a few millimeters; a thick isthmus is where goiters and the occasional isthmic nodule hide.
• Parenchymal echotexture — homogeneous, or heterogeneous with the increased vascularity and micronodular pattern of chronic lymphocytic thyroiditis. Say thyroiditis when you see it, because a sea of pseudonodules changes how much weight any single "nodule" deserves.
• One line on the neck: abnormal cervical lymph nodes, yes or no. A nodule with a suspicious ipsilateral node is a different conversation than the same nodule alone, and the lateral neck is part of the exam whether or not the technologist labeled images for it.

The five features, in prose

ACR TI-RADS scores exactly five things. Dictate each one explicitly for every reportable nodule, in the same order every time, and the report becomes self-scoring:

1. Composition — cystic, spongiform, mixed cystic and solid, or solid. "Spongiform" is a strong word: it means more than half the nodule is made of tiny cystic spaces, and it essentially ends the workup. Do not reach for it loosely.
2. Echogenicity — anechoic, hyperechoic, isoechoic, hypoechoic, or very hypoechoic (darker than the strap muscles). The comparison standard is the adjacent thyroid parenchyma; in a thyroiditis gland, say what you compared against.
3. Shape — wider-than-tall or taller-than-wide, measured on the transverse image. Taller-than-wide is the single heaviest feature in the system because benign nodules grow along tissue planes and cancers grow across them. If it's taller than wide, say so in those words.
4. Margin — smooth, ill-defined, lobulated, irregular, or with extrathyroidal extension. Ill-defined scores nothing; irregular scores. The distinction is whether you can't see the edge or the edge itself is wrong. Frank extension through the capsule is a surgical finding — describe what it touches.
5. Echogenic foci — none, large comet-tail artifacts, macrocalcifications, peripheral (rim) calcification, or punctate echogenic foci. Punctate foci are the sonographic correlate everyone worries about; comet-tail artifacts (big ones, in cystic components) are reassuring colloid. Don't let the two blur into "internal echogenicities."

Dictate the sum and the level — "TR4, 5 points" — and let the ACR TI-RADS calculator be the thing you verify against, not the thing you depend on. If you can't reconstruct the score from your own description, the description is incomplete.

FNA thresholds: the numbers that decide

The level only matters paired with the maximum dimension. The widely taught ACR thresholds:

• TR1, TR2 — no FNA, no follow-up, regardless of size.
• TR3 — FNA at ≥ 2.5 cm; follow-up ultrasound at ≥ 1.5 cm.
• TR4 — FNA at ≥ 1.5 cm; follow-up at ≥ 1.0 cm.
• TR5 — FNA at ≥ 1.0 cm; follow-up at ≥ 0.5 cm.

The discipline this buys you: a 9 mm TR5 nodule gets surveillance, not a needle, and your impression should say that plainly rather than hedging toward FNA because the features are ugly. Conversely a 3 cm TR3 nodule meets criteria even though nothing about it looks scary. Recommendations come from the table, not from vibes. Report the nodules that matter — the highest-scoring ones, conventionally up to four — rather than cataloguing every 4 mm colloid cyst in a multinodular gland; a ten-nodule inventory buries the one that needs the needle.

Surveillance: what counts as growth

When the recommendation is follow-up rather than FNA, the next report needs something to compare against — which is why the three dimensions and the image number are not clerical niceties. The widely taught ACR definition of meaningful growth is a 20% increase in at least two dimensions with a minimum 2 mm increase, or a 50% increase in volume; millimeter-level wobble between operators is noise, and calling it growth sends a stable nodule to the needle. On the follow-up study, re-score the features too — a nodule that crosses into a higher TI-RADS level is a new conversation even at a stable size, and your comparison line should say which happened: bigger, uglier, both, or neither.

Describe position for the person holding the needle

The FNA operator will find your nodule with a probe, not with your report's prose — unless the prose is good. Every reportable nodule gets:

• Side and third: right or left lobe; upper, mid, or lower pole — or isthmus.
• Depth and plane: anterior or posterior, medial or lateral.
• Three dimensions, with the convention that the first number is the one you'd follow over time.
• An image reference ("image 14") so nobody re-derives which nodule was scored.

"Right mid-pole, posterior and medial, abutting the trachea, image 14" tells the proceduralist the approach angle and the hazard before the probe touches skin. A posterior nodule against the carotid is a different stick than an anterior subcapsular one, and the report is where that planning starts.

The phrases attendings correct

• "Heterogeneous nodule." Not a TI-RADS word. Composition, echogenicity, shape, margin, foci — pick from the menu.
• "Too small to characterize." A thyroid nodule is never too small to characterize on ultrasound; it may be too small to matter, which is what the size thresholds are for.
• "Follow-up as clinically indicated." The indication is the table. Give the interval or give the needle.
• Scoring the gland, not the nodule. Thyroiditis background does not add points; punctate foci in a pseudonodule do not exist. When the gland is a minefield, say so and score only discrete nodules.

A model report

EXAMINATION: Ultrasound thyroid
HISTORY: 41-year-old female, palpable right neck fullness.
TECHNIQUE: Grayscale and color Doppler evaluation of the
thyroid gland and cervical soft tissues.
COMPARISON: None.
FINDINGS:
Right lobe: 5.2 x 1.8 x 1.9 cm. A solid, hypoechoic
nodule in the mid pole, posterior and medial, abutting
but not deforming the tracheal margin, measuring
1.6 x 1.2 x 1.1 cm (image 14). Wider than tall. Smooth
margin. No echogenic foci. ACR TI-RADS: solid (2),
hypoechoic (2), wider-than-tall (0), smooth margin (0),
no echogenic foci (0) = 4 points, TR4.
Left lobe: 4.8 x 1.6 x 1.7 cm. No discrete nodule.
Isthmus: 3 mm. Parenchyma: homogeneous.
Lymph nodes: No abnormal cervical lymph nodes.
IMPRESSION:
1. 1.6 cm TR4 nodule, right mid pole — meets ACR TI-RADS
   size threshold; FNA recommended.
2. No suspicious cervical lymphadenopathy.

Note what the model does: the five features appear in order with their points, the position reads like directions, and the impression converts the score into a verb. If you want the normal-study scaffold to build this discipline on, start from our normal thyroid ultrasound template. And if you'd rather watch a scratch dictation become exactly this report — features scored, thresholds applied, image numbers preserved — there's a live thyroid case in the demo, no signup, thirty seconds.