This review concluded that predominant internal flow seen on colour Doppler ultrasound was associated with malignancy of thyroid follicular neoplasms; absence of internal flow or predominantly peripheral flow indicated low probability of malignancy. The small number and relatively poor quality of included studies and lack of reporting of analytical methods mean that these conclusions should be interpreted cautiously.

To assess the diagnostic accuracy of colour Doppler ultrasonography for identifying malignancy in thyroid follicular neoplasms.

MEDLINE, Excerpta Medica, Web of Science, and the Latin American and Caribbean Health Sciences Literature database were searched. Search terms, but not search dates, were reported. Bibliographies of retrieved articles were screened for additional studies. No language restrictions were applied.

Studies that assessed colour Doppler ultrasonography for differentiation between benign and malignant thyroid follicular neoplasms, and that used histology of surgical biopsy as the reference standard to confirm diagnosis, were eligible for inclusion. Eligible studies were required to a perform colour Doppler ultrasonography evaluation before biopsy and fully describe flow patterns, differentiating at least two types of intra-nodular flow.

Most of the included studies were conducted in Japan. Included diagnostic data reported the detection of follicular adenomas, adenomatous goitres, follicular carcinomas, papillary carcinomas, follicular variants of papillary carcinomas, and adenomatous nodules; data on papillary carcinomas were excluded for one study.

Studies were independently assessed for inclusion by at least two reviewers and disagreements were resolved by consensus.

The methodological quality of included studies was assessed using seven items from the QUADAS (Quality Assessment of Diagnostic Accuracy Studies) tool including: inclusion of a representative spectrum of participants; use of an acceptable reference standard; acceptable delay between tests; avoidance of partial verification bias; avoidance of differential verification bias; index test results interpreted blind to reference standard; and reference standard interpreted blind to index test.

Three reviewers independently assessed studies and disagreements were resolved by consensus.

Data to populate 2x2 contingency tables were independently extracted by three reviewers and disagreements were resolved by consensus. Data were extracted for positive test of predominant internal flow, and for positive test of any internal flow. Sensitivity and specificity values, with 95% confidence intervals (CIs), were calculated for each data set.

Sensitivity and specificity estimates for individual included studies were presented in forest plots. The authors also reported "overall" sensitivity and specificity, positive and negative likelihood ratios (LRs), and positive and negative predictive values, but the methods used to calculate these values were not reported.

Four diagnostic cohort studies that recruited consecutive series of patients (457 thyroid follicular neoplasms) were included in the review. The overall prevalence of carcinomas was 14.7%.

The interval between index test and reference standard, procedure used for the reference standard and blinding of interpreters of index test and reference standard were not reported by any of the included studies. The remaining three quality criteria (representative spectrum of patients, avoidance of differential and partial verification bias) were met by all studies.

When predominant internal flow was used as the diagnostic criterion (four studies), the overall sensitivity was 85% (95% CI 74 to 93) and the overall specificity was 86% (95% CI 82 to 89). The positive likelihood ratio was 6.07 and the negative likelihood ratio was 0.18. At the prevalence reported in this study, the positive predictive value of colour Doppler ultrasonography for carcinoma would be 51% and the negative predictive value 97%.

When any internal flow was used as the diagnostic criterion (three studies), the overall sensitivity was 96% (95% CI 88 to 100) and the overall specificity was 14% (95% CI 11 to 18). For these data, the positive predictive value of colour Doppler ultrasonography would be 15% and the negative predictive value 96%.

Predominant internal flow seen on colour Doppler ultrasonography was associated with malignancy of thyroid follicular neoplasms. Absence of internal flow or predominantly peripheral flow indicated a low probability of thyroid follicular neoplasm malignancy.

The objective of the review was clearly stated. Appropriate inclusion criteria were defined. A number of sources were searched for relevant studies without apparent restrictions. Although the search terms were reported, the dates of the searches were unclear. Measures were taken throughout the review process to minimise error and/or bias.

The methodological quality of included studies was assessed and reported. Summary estimates of tests performance were reported, but their validity was unclear as no details of the analyses were reported.

Overall, the authors' conclusions should be interpreted cautiously, given the small number and relatively poor quality of included studies and lack of reporting of analytical methods.

Practice: The authors made no specific recommendations for practice.

Research: The authors stated that further prospective studies assessing the accuracy of predominant internal flow to identify thyroid follicular neoplasm malignancy are needed to corroborate their findings. They also stated that future evaluations of colour Doppler ultrasonography should describe more than the internal flow in thyroid nodules.

Not stated.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.