RET/PTC1 and RET/PTC3 are the markers for papillary thyroid carcinoma. Their reported prevalence varies broadly. Nonrearranged c-RET has also been detected in a variable proportion of papillary carcinomas. The published data suggest that a wide range in expression levels may contribute to the different frequency of c-RET and, particularly, of RET/PTC detection. However, quantitative expression analysis has never been systematically carried out. We have analyzed by real-time RT-PCR 25 papillary carcinoma and 12 normal thyroid samples for RET/PTC1, RET/PTC3 and for RET exons 10-11 and 12-13, which are adjacent to the rearrangement site. The variability in mRNA levels was marked and four carcinoma groups were identified: one lacking RET/ PTC rearrangement with balanced RET exon levels similar to those of the normal samples (7/25 cases, 28%), the second (6/25 cases, 24%) with balanced RET expression and very low levels of RET/PTC1, the third with unbalanced RET exons 10-11 and 12-13 expression, high RET/PTC1 levels but no RET/PTC3 (7/25 cases, 28%), and the fourth with unbalanced RET expression, high RET/PTC1 levels and low levels of RET/PTC3 (5/25 cases, 20%). Papillary carcinomas with high RET/PTC1 expression showed an association trend for large tumor size (P ¼ 0.063). Our results indicate that the variability in c-RET and RET/PTC mRNA levels contributes to the apparent inconsistencies in their reported detection rates and should be taken into account not only for diagnostic purposes but also to better understand the role of c-RET activation in thyroid tumorigenesis. Oncogenic c-RET activation in thyroid tumors composed of follicular cells is the result of chromosomal rearrangements resulting in the fusion of the RET tyrosine-kinase (RET-TK) domain to the 5 0 -terminal region of heterologous genes. The rearrangements are a molecular marker for papillary thyroid carcinoma as their very name, RET/PTC for papillary thyroid carcinoma, implies, 1 and consist of balanced inversions or translocations that involve the 3.0 kb intron 11 of c-RET. To date, at least 16 chimeric mRNAs affecting 11 different genes have been reported (Saenko et al, 2 reviewed in Tallini and Asa 3 ), of which RET/PTC1 (consisting of the fusion of RET with H4) and RET/PTC3 (consisting of the fusion of RET with RFG/ELE1) are by far the most common. 3 Their prevalence varies broadly from zero to more than 60% in nonradiation-associated cases 3 and is close to 90% in some series of papillary carcinomas from the Chernobyl area diagnosed after the 1986 nuclear disaster. 4 While the high prevalence of RET/PTC in radiation-associated thyroid tumors is consistent with misrepair of radiationinduced double-strand DNA breaks, 5,6 the high variability in the prevalence of RET/PTC in sporadic tumors has no specific explanation. In addition to rearranged RET forms, c-RET expression has also been identified in a highly variable proportion of papillary carcinoma samples. Although its signifi-