Dynamic Optical Contrast Imaging for the Pathological Diagnosis of Thyroid Cancer.

[OBJECTIVE] Initial biopsy specimens in suspected thyroid cancers may be nondiagnostic. Dynamic optical contrast imaging (DOCI) leverages endogenous fluorophore autofluorescence to examine the structural and metabolic properties of tissues. This preliminary study demonstrates DOCI's ability to differentiate between thyroid cancer histological subtypes.

[STUDY DESIGN] In vitro study of human thyroid tissue.

[SETTING] Single tertiary referral center.

[METHODS] Tissues from patients who underwent thyroidectomy for papillary (PTC), follicular (FTC), or anaplastic (ATC) thyroid cancer were included. Excised specimens underwent pathology review by a head and neck-trained pathologist. Corresponding unstained slides from identified regions of interest and normal thyroid tissue sections were imaged with DOCI. DOCI values of PTC, FTC, ATC, and normal thyroid tissue were compared.

[RESULTS] A total of 74 specimens (20 PTC, 20 FTC, 14 ATC, and 20 normal thyroid tissue samples) were collected and imaged with DOCI. In the spectral filters capturing collagen autofluorescence (408-425 nm), PTC displayed the highest DOCI signal, followed by ATC, normal thyroid tissue, and FTC (all comparisons P < .05). Significantly elevated DOCI values were observed in PTC versus ATC, and PTC versus normal thyroid in the spectral channels of nicotinamide adenine dinucleotide + hydrogen (481-505 nm) and flavin adenine dinucleotide (555-585 nm) autofluorescence (all P < .05). A random forest classifier trained on the data set to delineate cancer versus normal thyroid tissue demonstrated an accuracy of 94.6%.

[CONCLUSION] DOCI generates unique spectral profiles that delineate FTC, PTC, ATC, and normal thyroid tissue. By improving diagnostic accuracy, DOCI could reduce indeterminate results and overtreatment, especially in biopsy specimens.