Prostate cancer is the fifth most common cause of cancer death in men. In 2020, more than 375 thousand deaths attributed to this disease were reported. Nevertheless, these numbers could be drastically reduced if key procedures for early diagnosis such as digital rectal examination are used. However, despite its benefits, low participation in digital rectal examination remains a significant burden due to masculinity beliefs. An alternative to overcome this difficulty is the use of ultrasound scanning and deep learning methods for its processing. Other procedures used for detecting this type of cancer with high probability are the Prostate-Specific Antigen (PSA) and the International Prostate Symptom Score (IPSS). Indeed, using such procedures combined will allow higher accuracy in the diagnosis. Therefore, we propose ProstaTest, a system for non-invasive detection of prostate cancer based on IPSS, PSA and prostate ultrasound scanning applying Deep Learning, which altogether allows obtaining an automatic diagnosis with high accuracy. ProstaTest, with 48 ultrasound scans and 23 medical records, shows an accuracy of 96% in detecting prostate inflammation and 95.65% in diagnosing prostate cancer.
|Title of host publication||Proceedings of the 2021 IEEE Sciences and Humanities International Research Conference, SHIRCON 2021|
|Publisher||Institute of Electrical and Electronics Engineers Inc.|
|State||Published - 2021|
|Event||5th IEEE Sciences and Humanities International Research Conference, SHIRCON 2021 - Lima, Peru|
Duration: 17 Nov 2021 → 19 Nov 2021
|Name||Proceedings of the 2021 IEEE Sciences and Humanities International Research Conference, SHIRCON 2021|
|Conference||5th IEEE Sciences and Humanities International Research Conference, SHIRCON 2021|
|Period||17/11/21 → 19/11/21|
Bibliographical noteFunding Information:
ACKNOWLEDGMENT The authors would like to thank the institutions that provided the medical records and ultrasound scans used in this study as well as the specialists who participated in the surveys. This research was partially supported by Universidad Peruana de Ciencias Aplicadas (UPC), Project B060-2021.
© 2021 IEEE.
- deep learning
- image recognition
- prostate cancer
- ultrasound scans