A bibliometric analysis of use of Machine learning and Artificial
intelligence in Prostate Cancer Detection
Abstract
Objectives: Prostate cancer is one of the most common cancers worldwide
in men, with a huge geographical variation both in incidence and
mortality. Whereas, the incidence is higher in developed countries,
mortality is higher in developing countries. The reasons for high
mortality in these countries include variation in practice leading to
early diagnosis. Artificial Intelligence (AI) and Machine learning (ML)
are increasingly being used to improve the diagnostic accuracy of
prostate cancer. We interrogated the published literature to review the
usage of AI and ML in the diagnosis of prostate cancer. Methods:
Research databases such as SCOPUS, Web of Science (WoS), and Google
Scholar were searched to identify articles related to AI/ML in the
diagnosis and management of prostate cancer. Key-words included
(“prostate” AND “cancer”), (“machine” AND (“learn” OR
“learning”)) OR (“artificial” AND (“intelligence” OR
“intelligent”)). Results Using a screening criterion, 293 reviewed
research papers were identified. The two most consistent themes were
predictive modeling and application of AI/ ML tools for cancer grading
and radiomics. AI and ML enhance the diagnostic accuracy by reducing the
inter-individual variation in Gleason’s scoring, and complimenting the
interpretation of multiparametric magnetic resonance imaging (mpMRI). A
few publications reported the use of AI/ML tools by combining
histopathology and MRI signals. Conclusions: AI and ML can improve the
diagnostic accuracy of prostate cancer. Literature is beginning to
emerge suggesting to use a combination of demographic features, clinical
data, serological markers, pathological grading and radiological
factors, and genomic data, to propose accurate non-invasive diagnosis of
clinically significant prostate cancer.