Prostate cancer ranks as the second most prevalent cancer in men, posing a significant public health concern. While MRI scans are vital for early detection, the increased demand for scans has outpaced the growth of radiology experts. This has resulted in diagnostic delays and inconsistent interpretations, with only a minority of the medical community adhering to PI-RADS 2.1 guidelines.
Introducing QP-Prostate, an AI-powered solution designed to streamline radiologists’ workflows. By automatically evaluating image quality, segmenting the prostate gland, and identifying suspicious lesions*, QP-Prostate empowers radiologists to deliver quicker and more accurate assessments, ultimately enhancing patient care.
Automatic lesion detection*
QP-Prostate automatically highlights regions of the prostate suspicious for clinically significant prostate cancer (csPCa). QP-Prostate provides a High (orange) and Moderate (yellow) risk level, according to the risk of csPCa calculated by the software.
QP-Prostate detects suspicious lesions in both the Peripheral and Transitional+Central zones.
QP-Prostate’s AI algorithm automatically identifies the prostate anatomy and segments the full prostate gland, as well as the seminal vesicles, to assist in PSA density calculations. The results can be directly exported for fusion biopsy procedures.
QP-Prostate performs automated difussion (ADC map, synthesized b-1400s/mm2 ) and perfusion analysis to extract clinically meaningful quantitative information from the MRI study.
Cloud to PACS integration
All analysis outputs are directly sent back to the hospital PACS for their visualization without disrupting radiology workflows.
Structured radiology reports
QP-Prostate creates structured reports including quantitative information following PI-RADSv2.1 guidelines. The solution automates prostate gland volume calculation powered by AI, and identifies quantitative imaging biomarkers, ensuring precise measurements for diagnosis.
A suite of enhanced diagnostic capabilities
1. AI-based automated lesion detection*
Our AI algorithm, trained with pathology outcomes as ground truth, is designed to efficiently detect clinically significant prostate cancer lesions using biparametric inputs (T2w, DWI and ADC), with plans to outperform competitors in detection rate and speed. These AI algorithms are intended to elevate radiologists’ diagnostic accuracy with automated detection of biopsy-proven, clinically significant prostate cancer lesions.
2. Precision in segmentation
QP-Prostate’s AI algorithm, based on Convolutional Neural Networks (CNNs), automatically segments the prostate gland with market-leading performance (88% DSC1). It segments three key subregions (Peripheral, Transitional+Central zones, and Seminal Vesicles), incudes PI-RADS2.1 regions, and computes the prostate volume, facilitating PSA density calculations and fusion biopsy planning.
3. Quantitative diffusion and perfusion data
QP-Prostate automatically verifies MRI acquisition protocol according to PI-RADSv2.1 guidelines, ensuring that radiologists work with high-quality MRI examinations from the start.
QP-Prostate provides spatially registered, motion corrected, synthesized b-1400s/mm2 DWI and ADC map series, offering rich quantitative diffusion information. This dataset empowers radiologists to analyze potentially cancerous lesions with confidence. For mpMRI studies, perfusion analysis (Ktrans, Kep, Ve maps) are automatically generated.
4. Cloud to PACS integration
Ensure consistent and seamless incorporation of all outputs from QP-Prostate into your hospital’s PACS with no disruptions to radiologists’ workflow.
5. Structured radiology reports
QP-Prostate creates structured reports according to PI-RADSv2.1. The solution provides prostate volume and dimensions, and allows the radiologist to mark all relevant lesions and score them as per PI-RADSv2.1 guidelines.
The product will be enhanced with predictive capabilities in upcoming versions, which are currently undergoing clinical studies for imaging-based prediction of biochemical relapse. Preliminary results show that combining MRI with clinical data predicts a 10-year biochemical recurrence with an area under the curve (AUC) of 0.84 to 0.872.
Changing the narrative in prostate diagnostics
– Jimenez-Pastor A, et al. Eur Radiol. 2023;33(7):5087-5096.
– Sánchez Iglesias Á, et al. Cancers (Basel). 2023;15(16):4163.
*In the US: Lesion detection functionality is under clinical investigation and not available for sale.