Diagnostic accuracy
Can ultrasound scores sort unclear thyroid biopsies?
Open access · cc by · source: Europe PMC
For thyroid nodules whose needle biopsy was inconclusive, combining the biopsy category with a standard ultrasound risk score picked out low-risk nodules that were almost always benign.
Study at a glance
- Design
- Cross-sectional — Retrospective diagnostic-accuracy study: blinded re-scoring of pre-biopsy ultrasound images compared with post-thyroidectomy histology as reference standard
- N
- N=140 · 140 cytologically indeterminate thyroid nodules in 139 patients who had surgery, drawn from all thyroid biopsies at two pathology centres
- Population
- Adults (mostly women) in Paraná, Brazil (a university hospital plus one private centre), with Bethesda III, IV or V thyroid nodule cytology who underwent thyroidectomy, 2012-2016
- Outcome
- Malignancy on final histopathology; sensitivity, specificity, PPV, NPV and accuracy of ACR TI-RADS and 2015 ATA ultrasound categories alone and combined with Bethesda category
Structured fields used in claim comparison tables when every cited study has a complete layer.
Key findings
Of the 140 nodules, 66 (47.1%) were malignant, and malignancy rose with cytology category (15.9% for Bethesda III, 40.4% for IV, 86.4% for V) and with ultrasound suspicion. On their own, both ultrasound systems had modest sensitivity (69.7%) with specificity of 81.1% (TI-RADS) and 86.5% (ATA), and the two systems agreed closely (kappa 0.91). Combining them with cytology worked better: Bethesda III nodules with a low-risk ultrasound were malignant only 5.9% (TI-RADS) or 5.7% (ATA) of the time, giving negative predictive values around 94%, while Bethesda IV/V nodules with high-risk ultrasound were mostly cancers.
Methodology
The team searched biopsy records from two pathology centres in Brazil for thyroid nodules with indeterminate cytology (Bethesda III, IV or V) that also had stored ultrasound images and a surgical pathology result. A radiologist blinded to cytology and histology re-scored each nodule's ultrasound using the 2017 ACR TI-RADS points system and the 2015 American Thyroid Association (ATA) patterns. Nodules were labelled favourable or unfavourable by ultrasound and by cytology, and these labels were compared with whether surgery found cancer.
Limitations
Only nodules that went to surgery and had retrievable images could be included, so the sample is heavily filtered and malignancy (47%) is much higher than in all indeterminate nodules, which inflates predictive values; the authors note the high NPV may not hold where cancer risk differs. A single radiologist re-read static stored images rather than scoring live scans, and reader agreement between radiologists was not tested. The study is retrospective, from one region, and lacked data such as TSH and thyroid autoimmunity.
How this study connects
Role on claims
Each row is a claim on a concept or method page where this paper supports, challenges, or qualifies the statement. Roles are hand-checked — not a model guess.
Not yet placed on a claim. This paper has study layers, but no concept page yet cites it as support, challenge, or qualifier.
Related papers in this topic
Same topic cluster — not a recommendation engine.