Software-Based Analysis Aids Ocular Surface Squamous Neoplasia Diagnosis

Close up of the conjunctival squamous cell carcinoma.
The technology relies on analysis of a few biomarkers.

Clinical markers, including HIV seropositivity and the presence of feeder vessels are key in identifying and differentiating invasive squamous cell carcinoma (iSCC) from conjunctival intraepithelial neoplasia (CIN), according to research published in International Ophthalmology. Image processing software can be used to conduct objective analyses of ocular surface squamous neoplasia (OSSN), aiding in diagnosis. 

Researchers conducted a retrospective case control study to assess the morphological parameters that aid in clinical differentiation of CIN from iSCC. Additionally, researchers aimed to explore the potential utility of imaging processing software in objectively assessing OSSN. 

Participants were patients at the Operation Eyesight Universal institute for Eye Cancer in India. Clinical records of all participants were retrospectively reviewed for sociodemographic characteristics, seropositive status for HIV, and clinical tumor characteristics. 

In total, 108 OSSN lesions from 107 patients were included (CIN n=75 eyes; iSCC n=33 eyes). At baseline, there were no significant differences in demographic or socioeconomic markers or symptoms; however, the iSCC group had a significantly higher proportion of patients who were seropositive (odds ratio [OR], 11.28; 95% CI, 4.21-30.25). Additionally, 4 patients in the CIN group and 1 in the iSCC group had xeroderma pigmentosum. 

Investigators found that patients in the iSCC group had significantly higher maximum diameter compared with the CIN group (median, 6.9 mm vs 5 mm; interquartile range [IQR], 6 mm to 9 mm vs 3.5mm to 7.2 mm). Measurements for height and area were also higher compared with CIN. 

In terms of tumor morphology, the iSCC group had a “significantly higher” proportion of nodular OSSNs (OR, 3.30; 95% CI, 1.40-7.74; P =.01) and a significantly lesser proportion of gelatinous morphology (OR, 0.09; 95% CI, 0.01-0.72; P =.01). Feeder vessel presence was significantly higher compared with CIN. Other tumor features — such as leukoplakic and papilliform morphology, quadrant distribution, tumor epicenter, surface keratin, and pigmentation, among others — were comparable between groups. 

Results of a univariate logistic regression analysis identified HIV seropositivity, maximum diameter, perpendicular to maximum diameter, height, nodular morphology, gelatinous morphology, and feeder vessel presence as significant predictors of iSCC. Using these factors as independent variables in a multiple logistic regression analysis, investigators found that the only predictor of iSCC was HIV seropositivity (OR, 13.33 ± 8.35; 95% CI, 3.90-45.33; P <.0001). 

Study limitations included those inherent to retrospective studies and the small sample size. 

Due to the nature of OSSN, histopathology has remained the standard to distinguish CIN from iSCC. Investigators confirmed that large, thick, nongelatinous, nodular OSSN lesions with feeder vessels should raise “a high index of suspicion” for iSCC, and prompt HIV screening. 

“Objective analysis of ocular surface lesions using automated software is a simple, feasible, and objective method for the assessment of ocular surface lesions,” the researchers concluded. “Moving forward, further advances in medical terminology and innovation can bring forth an automated and user-friendly software which can be incorporated in the slit lamp digital camera to analyze images at the time of capture itself, helping ophthalmologists in clinical decision making.” 


Vempuluru VS, Kapoor AG, Kaliki S, Jajapuram SD, Mohamed A, Mishra DK. Comparative evaluation of clinical characteristics of biopsy-proven conjunctival intraepithelial neoplasia and invasive squamous cell carcinoma using image processing software programs. Int J Ophthalmol. Published online January 4, 2021. doi:10.1007/s10792-020-01687-9