Discrepancy between ultrasound and final pathologic measurements in thyroid cancer

The size of the malignancy in well-differentiated thyroid carcinoma (WDTC) is an important factor in patient risk stratification. Final pathologic measurements are currently used for treatment decisions. Measurements are made by both ultrasonography and on final pathology and should, theoretically, agree with each other.

To evaluate the agreement between ultrasound and final pathologic measurements in thyroid nodules and WDTC.

A retrospective chart review of all total thyroidectomies between 1998 and 2007 was completed by four different surgeons at one centre. Thyroid specimens with a solitary nodule and complete ultrasound and pathologic data were included. A comparison of discrepancy and concordance of the largest dimension of the nodule by ultrasound and pathologic measurement was made.

A total of 263 thyroid specimens with a solitary nodule were identified. Of these, 68 were found to harbour WDTC. A median underestimation by pathology of 0.54 cm was present in WDTC, averaging 20% of the nodule size by pathology. This trend was consistent with subgroup analysis for size, pathology, and nodularity. An underestimation by pathology of 0.4 cm was present for papillary carcinoma under 1 cm. Grouping nodules by size, a concordance of 80% was present for nodules under 1 cm. Statistical testing revealed weak correlation and a kappa of approximately 0.3 for agreement. Weighted kappa was approximately 0.5 for agreement.

The recent advances in ultrasonography and the constraints of pathology preparation allow an argument to be made for investigation of using ultrasound measurements as an adjunct in making treatment decisions.