Table 1.

Published studies analyzing molecular transcriptomic or proteomic specificities of radiation-induced thyroid cancers.

	Study	Port et al.; 2007	Detours et al.; 2007	Stein et al.; 2010 1	Ugolin et al.; 2,3	Ory et al.; 2011 3	Boltze et al.; 2009
	Exposure	Post-Chernobyl	Post-Chernobyl	Post-Chernobyl	Post-Chernobyl	Post-radiotherapy	Post-Chernobyl
Radiation-induced	Tumor set	11 PTC 6	12 PTC	10 PTC	Learning set: 6 PTC	Learning set: 7 rPTC, 7rFTA	86 PTC
		3 males, 8 females	4 males, 8 females	Half male and half female	3 males, 3 females	4 males, 10 females	40 males, 46 females
	Age at IR	/	1b–16 years(M = 8.6 years)	2 monthb–14 years (M = 6 years)	10b–16 years (M = 12.7 years)	3b–14 years (M = 8.6 years)	3b–23 years (M = 11.8 years)
	Age at diagnosis	15b–22 years (M = 18 years)	16b–33 years (M = 24 years)	14b–28 years (M = 20 years)	27b–33 years (M = 29.2 years)	20b–56 years (M = 35.1 years)	12b–28 years (M = 18.6 years)
	Latency	Up to 15 years after 1986	15b–17 years (M = 16 years)	14b–16 years (M = 13.6 years)	16b–17 years (M = 16.5 years)	11b–48 years (M = 26.5 years)	Up to 15 years after 1986 (mean = 6.8 years)
	Dosimetry	90% at 0.15-1Gy 4	/	/	/	12b–42.5 Gy (M = 14.1 Gy)	90% at 0.15-1Gy 4
	Histology	PTC	8 PTC, 3 FVPTC, 1 PTC	PTC	? 7	7 FTA; 6 PTC; 1 FVPTC	PTC
	Mutations	2 PTC1; PTC3 6	5 BRAF; 5 PTC	1 PTC; 1 PTC3	4 BRAF; 1 RET/PTC	1 BRAF; 1 RAS
	Others	T2N0M0 to T4N1M1	/	/	/	5 with chemotherapy
Sporadic	Tumor set	41 PTC	14 PTC 7	20 PTC from He et al. study 5	Learning set: 7 PTC	Learning set: 7 sPTC, 7sFTA	91 PTC
		19 males, 22 females	5 males, 9 females	8 males, 12 females	4 males, 3 females	5 males, 9 females	49 males, 42 females
	Age at diagnosis	15b–83years (M = 60 years)	29b–68 years (M = 47 years)	13b–65 years (M = 44.4 years)	29b–38 years (M = 34.6 years)	21b–63 years (M = 37.6 years)	15b–83 years (M = 50.1 years)
	Histology	PTC 6	9 PTC, 4 FVPTC, 1 tPTC	14 PTC; 5 FVPTC; 1 HCC7	/ 7	7 FTA; 5 PTC; 2 FVPTC	PTC
	Mutations	5 PTC1	5/14 BRAF; 3/14 RET/PTC	/	2 BRAF; 1 RET/PTC	4 BRAF; 2 RAS, 1 PTC1, 1 PTC3
	Others	T1N0-1M0 (n = 26)  to T3N1M0	/	/		None with chemotherapy
	Arrays	Human genome survey microarray V2.0 (Applied Biosystems) (33,000 probes)	Human 1 cDNA Microarray slides (Agilent Technologies). (19,000 probes)	Affymetrix U133A Array (20,000 probes) (Stein et al.)	Dataset  retrieved from GEO (GSE3950)	Human 25K 50b–52mer oligo-microarrays (national genomic platform)	(Not relevant)
			Hybridized with normal matched tissue	Affymetrix U133 Plus 2.0 Array (50,000 probes) (He et al.) Hybridized with normal matched tissue		Hybridized with an internal reference (pool of normal thyroid tissues)
	Analysis	Identified 1300 genes up- or downregulated at least fivefold  (pool of 10 rPTC vs. pool of 10 sPTC)	(1) Several methods applied for tumor classification	Compared two sets of deregulated genes obtained separately: (1) post-Chernobyl PTC vs. normal tissues and; (2) sporadic PTC vs. normal tissues	106 genes discriminating signature identified by applying the EMts_PCA on the learning/training set	322-gene discriminating signature identified by applying the EMts_PCA to the learning/training set	Identification of protein markers by MALDI-TOF mass spectrophotometry
		Validation of 92 more deregulated genes in the full tumor series by RT-PCR	(2) Same methods applied for tumor classification by using a γ-irradiation vs. H2O2 lymphocyte response signature.	Retained the genes deregulated in post-Chernobyl PTC only	651 deregulated genes identified	1900 deregulated genes identified	20 candidate protein markers analyzed by immunochemistry
	Results	10 genes for complete separation of the groups (no validation on an independent tumor sets)	In both cases classification with error rate errors of 8 to 42% for sporadic tumors and 7 to 29% for post-Chernobyl tumors	Identified 177 deregulated genes unique to the radiation-induced tumors	Etiology prediction of the 13 remaining tumors using the 106 gene signature (1 unclassified, non misclassified)	Blind prediction of etiology of the 29 remaining tumors (tumors (13 rPTC or FTA; 16 sPTC FTA) (1 unclassified, 2 misclassified)	Combination of 6 of these markers separates the groups (no validation on an independent tumor sets)

PTC: Papillary thyroid carcinoma; FVPTC: PTC, follicular variant, tPTC: PTC, trabecular variant; PTCs: PTC, solid variant; HCC: Hurthle cell carcinoma (HCC with follicular and papillary features); FTA: follicular thyroid adenoma; rPTC, rFTA: radiation-induced PTC; FTA; sPTC, sFTA: sporadic PTC; FTA; RAS: mutation in NRAS, HRAS or KRAS gene; BRAF: V600E BRAF mutation; PTC (unspecified), PTC1, PTC3: RET/PTC rearrangement; M: mean; ¹ Transcriptome analysis was performed on 10 out 14 PTC tumors described in Stein et al. [32]. Clinical data are given for the full tumor set; ² From Detours et al. [29]. For the analysis described in Ugolin et al. [34], 6 out of 12 post-Chernobyl PTCs and 7 out 14 sporadic PTCs of the tumor set described by Detours et al. [29], were used as a learning/training set for signature identification, the remaining tumors were used as testing set. Clinical data are given for the learning/training set; ³ Clinical data are given for the learning/training set; ⁴ Estimation from general dosimetry data; ⁵ From the He et al. study, 2005 (GSE3467); ⁶ No indication of the 10 tumors used for pool; ⁷ No indication of the precise histology by tumor.