Data-driven overdiagnosis definitions: A scoping review

Validation and three years of clinical experience in using an artificial intelligence algorithm as a second read system for prostate cancer diagnosis-real-world experience

Background

Prostate cancer ranks as the most frequently diagnosed cancer in men in the USA, with significant mortality rates. Early detection is pivotal for optimal patient outcomes, providing increased treatment options and potentially less invasive interventions. There remain significant challenges in prostate cancer histopathology, including the potential for missed diagnoses due to pathologist variability and subjective interpretations.

Methods

To address these challenges, this study investigates the ability of artificial intelligence (AI) to enhance diagnostic accuracy. The Galen™ Prostate AI algorithm was validated on a cohort of Puerto Rican men to demonstrate its efficacy in cancer detection and Gleason grading. Subsequently, the AI algorithm was integrated into routine clinical practice during a 3-year period at a CLIA certified precision pathology laboratory.

Results

The Galen™ Prostate AI algorithm showed a 96.7% (95% CI 95.6-97.8) specificity and a 96.6% (95% CI 93.3-98.8) sensitivity for prostate cancer detection and 82.1% specificity (95% CI 73.9-88.5) and 81.1% sensitivity (95% CI 73.7-87.2) for distinction of Gleason Grade Group 1 from Grade Group 2+. The subsequent AI integration into routine clinical use examined prostate cancer diagnoses on >122,000 slides and 9200 cases over 3 years and had an overall AI Impact ™ factor of 1.8%.

Conclusions

The potential of AI to be a powerful, reliable, and effective diagnostic tool for pathologists is highlighted, while the AI Impact™ in a real-world setting demonstrates the ability of AI to standardize prostate cancer diagnosis at a high level of performance across pathologists.

Artificial Intelligence in Prostate Cancer Diagnosis: “Synergy-Net” in Campania FESR-POR (European Fund of Regional Development—Regional Operative Program) Research Project

Background: The diagnosis of prostate cancer can only be obtained following the analysis of the tissue taken by means of a biopsy. Given the position of the organ, the biopsy is typically assisted by ultrasound images and the procedure consists of taking different portions of tissue from different areas, according to a map well-defined by international standards. Given the invasiveness of the procedure, the objective set within the Synergy-Net project is to analyze biomedical images in order to guide the operator on identifying the most suspicious tissues. Results: The dataset acquired by the Synergy-Net Platform at the “Vanvitelli” Urology Operating Unit is made up of a total of 350 outpatient services from which the diagnosis emerged on ultrasound, elastosonography, RNM, and biopsy of 50 prostate carcinomas which were then operated on. In the context of the Synergy-Net project, a new convolutional architecture was therefore created based on the U-Net paradigm, designed to perform a slice-by-slice segmentation in DCE-MRI of the prostate. The data processing with CNNs was carried out on a dataset of 37 patients, selected from the initial 50 for completeness and uniformity of the data, all affected by k-prostatic disease, using a tenfold cross-validation in order to obtain a statistically more significant estimate of the goodness of the results obtained. The performance metric used was the DICE coefficient. Conclusion: The results present a low intra-subject variability and a high inter-subject variability, with DICE values ranging between a minimum of 5.8% and a maximum of 60.3%. On average, a value of 35% is reported, considering the arithmetic mean of the dice achieved on all folds (macro-average).