Artificial Intelligence has potential for Prevention and early diagnosis of Cervical cancer, claims research

Cervical cancer is the primary cause of increased morbidity and mortality worldwide. Effective screening of precancerous lesions helps in preventing cervical cancer. Cervical cytology screening and colposcopy are vital tools to prevent cervical intraepithelial neoplasia (CIN) and cervical cancer. Recently, AI has been used for various medical image analysis tasks to identify various diseases. Previous literature has shown the magnificent diagnostic accuracy of AI in detecting CIN and cervical cancer. However, a meta-analysis of all the studies has not been performed. Hence, researchers conducted a systematic review and meta-analysis to examine the pooled accuracy, sensitivity, and specificity of AI-assisted cervical cytology screening and colposcopy for cervical intraepithelial neoplasia and cervical cancer screening.

Various databases such as PubMed, Embase, and Cochrane Library were used to collect data between January 1, 1986, and August 31, 2024. Various studies that investigated the sensitivity and specificity of AI-assisted cervical cytology screening and colposcopy for histologically verified cervical intraepithelial neoplasia and cervical cancer and a minimum of five cases were included. The performance of AI and experienced colonoscopists were used to evaluate various metrics along with the area under the receiver operating characteristic curve (AUROC), sensitivity, specificity, accuracy, positive predictive value (PPV), and negative predictive value (NPV), analyzed through random effect models. Subgroup analyses were done to compare diagnostic performance across developed and developing countries.

Findings:

• About 77 studies that met the inclusion criteria were included.
• The pooled diagnostic performance metrics for AI-assisted cervical cytology using Papanicolaou (Pap) smears were as follows:

accuracy	94% (95% CI 92–96);
sensitivity,	95% (95% CI 91–98);
specificity,	94% (95% CI 89–97);
positive predictive value (PPV),	88% (95% CI 78–96);
negative predictive value (NPV),	95% (95% CI 89–99)

• For AI-assisted cervical cytology using the ThinPrep cytologic test (TCT), the pooled metrics were:

accuracy,	90% (95% CI 85–94);
sensitivity,	97% (95% CI 95–99);
specificity,	94% (95% CI 85–98);
PPV,	84% (95% CI 64–98);
NPV,	96% (95% CI 94–98).

Thus, the study concluded that AI can be used as a potential source for early diagnosis and prevention of cervical cancer. AI demonstrates enhanced diagnostic accuracy and can be used as a reliable tool for diagnostic precision. The study underscores the importance of using AI as a cost-effective tool for cervical cancer screening, minimizing observer variability, and improving outcomes.

Further reading: Liu, Lei et al. Performance of artificial intelligence for diagnosing cervical intraepithelial neoplasia and cervical cancer: a systematic review and meta-analysis. eClinicalMedicine, Volume 80, 102992. Doi: 10.1016/j.eclinm.2024.102992.