Bridging the Experience Gap in Prostate Multiparametric Magnetic Resonance Imaging Using Artificial Intelligence: A Prospective Multi-Reader Comparison Study on Inter-Reader Agreement in PI-RADS v2.1 - Beyond the Abstract

Artificial Intelligence (AI) has the potential to improve the detection of prostate cancer (PCa) by standardizing imaging results and enhancing the interpretation of prostate Magnetic Resonance Imaging (MRI). Radiologists' expertise remains vital in accurately diagnosing and interpreting imaging results.

To this end, two systems have been introduced: the Prostate Imaging-Reporting and Data System (PI-RADS) version 2.1, which enhances reader performance, and the Prostate Imaging Quality system (PI-QUAL), which standardizes the quality control of MRI scans.^1-3

AI can assist radiologists in daily tasks such as lesion detection, image quality assessment, and risk stratification.^4-7 Further investigation is required to evaluate its impact on inexperienced radiologists and reporting times.

A prospective observational study was conducted using the commercially available software, Quantib® Prostate, to interpret multiparametric MRI (mpMRI) scans in patients suspected of having PCa and assess its impact on the inter-reader agreement between inexperienced and expert readers, reporting time, and agreement at different PI-QUAL scores and Grades of Confidence.

The study included 200 patients divided between four equal batches of 50. Scans were evaluated by an expert radiologist, acting as the reference standard, and four novice readers using PI-RADS v2.1 criteria without the AI-assisted software.⁶ The same cases were evaluated with the software by the novice readers after a 2-week interval, and three weeks were allowed between each batch for interactive case study teaching sessions. A final round of reader evaluations was conducted for the first batch.

The use of Quantib® improved inter-reader agreement in less experienced readers, while more experienced readers achieved higher diagnostic accuracy without it. The software led to longer reporting times for all readers, with uploading time being the most time-consuming step in the workflow, followed by segmentation and lesion identification. The findings indicate that Quantib® may be useful for less experienced readers and in cases where image quality and grade of confidence are suboptimal.

Further investigation is necessary to determine the effect of Quantib® on the potential problem of excessive dependence on software, as well as its impact on clinical outcomes measured against the reference standard of biopsy samples. Proficiency in using the software requires a period of adaptation, during which even experienced readers may encounter difficulty accommodating changes to their usual workflow.

Quantib® should not replace the local PACS but rather complement it. It can potentially standardize prostate MRI interpretation, enhance diagnostic accuracy, and improve agreement between readers, depending on the reader's expertise level. Such software can assist less experienced radiologists in detecting clinically significant lesions and improve prostate MRI interpretation. However, its use should be customized based on the reader's experience and the desired examination outcome. Radiologists require proper training to use this technology, and the difficulties of integrating it into daily clinical practice must be addressed.

References:

1. F. Giganti, C. Allen, M. Emberton, C.M. Moore, V. Kasivisvanathan, Prostate Imaging Quality (PI-QUAL): A New Quality Control Scoring System for Multiparametric Magnetic Resonance Imaging of the Prostate from the PRECISION trial, European Urology Oncology. 3 (2020) 615–619.
2. F. Giganti, V. Panebianco, C.M. Tempany, A.S. Purysko, Is Artificial Intelligence Replacing Our Radiology Stars in Prostate Magnetic Resonance Imaging? The Stars Do Not Look Big, But They Can Look Brighter, European Urology Open Science. 48 (2023) 12–13.
3. Turkbey, B., Rosenkrantz, A. B., Haider, M. A., Padhani, A. R., Villeirs, G., Macura, K. J., Tempany, C. M., Choyke, P. L., Cornud, F., Margolis, D. J., Thoeny, H. C., Verma, S., Barentsz, J., & Weinreb, J. C. (2019). Prostate Imaging Reporting and Data System Version 2.1: 2019 Update of Prostate Imaging Reporting and Data System Version 2. European urology, 76(3), 340–351.
4. B. Turkbey, M.A. Haider, Deep learning-based artificial intelligence applications in prostate MRI: brief summary, BJR. 95 (2022) 20210563.
5. T. Penzkofer, A.R. Padhani, B. Turkbey, M.A. Haider, H. Huisman, J. Walz, G. Salomon, I.G. Schoots, J. Richenberg, G. Villeirs, V. Panebianco, O. Rouviere, V.B. Logager, J. Barentsz, ESUR/ESUI position paper: developing artificial intelligence for precision diagnosis of prostate cancer using magnetic resonance imaging, Eur Radiol. (2021).
6. R. Suarez-Ibarrola, A. Sigle, M. Eklund, D. Eberli, A. Miernik, M. Benndorf, F. Bamberg, C. Gratzke, Artificial Intelligence in Magnetic Resonance Imaging–based Prostate Cancer Diagnosis: Where Do We Stand in 2021?, European Urology Focus. 8 (2022) 409–417.
7. Cipollari, S., Pecoraro, M., Forookhi, A., Laschena, L., Bicchetti, M., Messina, E., Lucciola, S., Catalano, C., & Panebianco, V. (2022). Biparametric prostate MRI: impact of a deep learning-based software and of quantitative ADC values on the inter-reader agreement of experienced and inexperienced readers. La Radiologia medica, 127(11), 1245–1253.
8. I.G. Schoots, J.O. Barentsz, L.K. Bittencourt, M.A. Haider, K.J. Macura, D.J.A. Margolis, C.M. Moore, A. Oto, V. Panebianco, M.M. Siddiqui, C. Tempany, B. Turkbey, G.M. Villeirs, J.C. Weinreb, A.R. Padhani, PI-RADS Committee Position on MRI Without Contrast Medium in Biopsy Naive Men with Suspected Prostate Cancer: A Narrative Review, American Journal of Roentgenology. (2020) AJR.20.24268.