Key points
Early diagnosis and treatment of vocal cord leukoplakia may prevent
progression and malignancy.
Lesions of vocal cord leukoplakia were classified as nonsurgical group
(NSG) and surgical group (SG) using pathology as a gold standard.
We applied deep learning (DL) with convolutional neural networks (CNNs)
to segment and classify narrow band imaging (NBI) and white light
imaging (WLI) of vocal cord leukoplakia .
The DL model could detect the lesions autonomously with average
intersection-over-union (IoU) value exceeded 70%. For classification,
the model was able to classify the surgical group (SG) by laryngoscope
with a sensitivity of 0.93 and specificity of 0.94 in WLI, and a
sensitivity of 0.99 and specificity of 0.97 in NBI. Moreover, this model
achieved a mean average precision (mAP) of 0.81 in WLI and 0.92 in NBI
with an IoU > 0.5.
The novel DL model shows promise in assisting in accurate diagnosis of
vocal cord leukoplakia from WLI and NBI.