The diagnosis and prognosis of follicular thyroid neoplasms are based on the identification of capsular and vascular invasion. Although conventional histology allows accurate classification in most cases, its inherent limitations in tissue sampling can result in misjudgment of both the presence and extent of invasion, and occasionally result in diagnostic inaccuracies. Consequently, unexpected tumor recurrence or overtreatment still represent significant clinical challenges. To mitigate these limitations, emerging diagnostic approaches are exploring advanced imaging modalities. X-ray 3D virtual histology has been reported to enable non-destructive and comprehensive sampling of the entire tumor volume embedded in formalin-fixed paraffin-embedded blocks. In thisstudy, the X-ray 3D virtual histology technique is first evaluated by classifying 99 follicular thyroid carcinomas and 31 follicular adenomas, achieving an accuracy of 89.2%. It is then applied to tissue blocks from five relapse cases that were postoperatively diagnosed as adenomas using conventional histology. Three of five tumors exhibited at least one unequivocal focus of vascular invasion, reinforcing that even a single well-defined focus portends significant risk of recurrence and distant metastasis. Although histological confirmation at this stage remained necessary, X-ray 3D virtual histology proved to be a valuable screening method.
Keywords: Digital pathology; Follicular thyroid neoplasms; Micro-CT; X-ray 3D virtual histology.
© 2025. The Author(s).