[Artificial intelligence aided measurement of cervical squamous epithelial thickness and its correlation with cervical precancerous lesions].

Objective: To study the thickness of cervical squamous epithelia and its correlation with cervical precancerous lesions. Methods: We selected 495 HE slides of 209 cervical biopsies from January 2020 to June 2020 in the Department of Pathology, the First and Seventh Medical Center of the PLA General Hospital, including 173 slides with low grade squamous intraepithelial lesion (LSIL) and 214 slides with high grade squamous intraepithelial lesion (HSIL). Artificial intelligence labeling software was used to assist in measuring the epithelial thickness of normal cervical squamous epithelium, LSIL and HSIL of each slide. The thickest, thinnest, and middle widths of epithelial thickness were measured, respectively. Average epithelial thickness was defined as the sum of the above three widths divided by 3. The correlation statistical analysis was performed by combining the data of age and pathological diagnosis. Results: The average thickness of normal cervical squamous mucosa was (245.83±91.40) µm, which was (222.42±81.22) µm and was (195.95±66.59) µm in LSIL and HISL epithelial respectively (F=27.09, P<0.01). The average cell layers of normal cervical squamous epithelium was (15.5±4.2) layers, which of LSIL was (14.8±4.8) layers, and that of HSIL was (15.8±4.8) layers. The differences among normal, LSIL and HSIL were not statistically significant (P>0.05). Further statistical analysis was stratified by age (≤30 years, 31-40 years, 41-50 years, 51-60 years, and >60 years), the results of Pearson correlation analysis showed that the thickness of normal cervical squamous epithelial gradually thinned with age (correlation coefficient r=-0.141 9, P<0.05), while LSIL and HSIL epithelial thickness had significant correlation with age (P>0.05). In the subgroup of ≤50 years old, the epithelial thickness of normal squamous epithelium was the thickest, followed by LSIL, and HSIL epithelial thickness was the thinnest. The differences were statistically significant (P<0.05). While in the subgroup of >50 years, the differences were not statistically significant (P>0.05). Conclusions: The cervical squamous epithelium gradually becomes thinner with the degree of precancerous lesions increasing among patients of ≤50 years old. However, after age of 50 years, with the onset of menopause, the normal mucosal epithelium is becoming atrophy, so that mucosal thickness is no longer correlated with the extent of the lesion. In addition, it is suggested that the cervical vinegar white test performance during colposcopy is related to the protein changes in the mucosal epithelial cells, but not directly related to the thickness of the epithelial layer.