Innovations of 2020: what is currently at the cutting edge?
Uro-radiology continues to evolve from anatomic to more functional assessment of disease processes. Recently, the introduction of radiomics has enabled the automated extraction and analysis of salient features from acquired images. Artificial intelligence (AI) and deep learning are profoundly shaping modern imaging, with major implications for medical practice. Soon, AI may speed up quantitative imaging by automating time-consuming tasks of measuring features and may be used as a tool to double-check the consistency of human reports, to prevent missed or wrong diagnoses.
The thrills of this year include applications of AI, the characterisation of renal tumours in kidney imaging, the new VI-RADS of the bladder and multiparametric imaging of the prostate, which has revolutionised the management of prostate cancer, and should now be performed before biopsy.
The state of the art of kidney image segmentation has improved greatly during the past five years, thanks to the progress accomplished in 2D automatic measurements and deep learning techniques. Future research is needed to address the issue of renal cortex segmentation in 3D. Research groups have demonstrated the potential role of multiparametric MRI to accurately differentiate the most common histologic types of renal cancer, which has significant prognostic implications. CT and MR texture analyses are opening up a new era and seem to be able to differentiate between indolent and more aggressive histology of renal tumours. Recently, the connection between histopathology imaging and proteomics in kidney cancer through machine learning has also been investigated, with promising results.
Regarding bladder imaging, the major advance is the development of the VI-RADS scoring system, by Panebianco et al. in 2018, to differentiate non-muscle from muscle-invasive bladder cancer, standardising its reporting. Recent validation studies have demonstrated its accuracy and reproducibility with good inter-reader agreement.
What the future holds: insight into MRI of the prostate
At the present time, the past utopian visions of MR-guided procedures, standardised scores and staging systems have become reality. Years of innovations have produced outstanding advances in diagnostic imaging and MR-guided interventional procedures.
Among all the works supporting the use of MRI in the diagnostics of prostate cancer, three robust 1A evidence level studies and one meta-analysis have played a pivotal role. The Precision clinical trial, the MRI First, the 4M validating pairing studies and the recently published Cochrane 2019 meta-analysis have shown the superiority of performing mpMRI followed by targeted biopsy in patients with a suspicion of prostate cancer: the ‘MRI pathway’ of the prostate.
In early 2019, the updated version of the PI-RADS score system (v. 2.1) was released. Inconsistencies and limitations from the previous version were addressed, particularly narrowing the criteria to define the PI-RADS 3 score. In the document, the role of the quantitative analysis of MRI functional sequences parameters has been addressed and approved to further stratify prostate cancer suspicion.
The success of the technique is heavily dependent on high-quality image acquisition and interpretation. Indeed, ‘expert hands’ are needed to acquire high quality MRI, minimise the number of equivocal cases with biopsy avoidance, standardise negative MRIs, reduce overdiagnosis and overtreatment and promote biopsy improvement and focal therapeutic approaches. All of which might be addressed by applying artificial intelligence techniques that might, among many goals, reduce the large variability and detect MRI silent tumours, with the final implication of improving the technical aspects of MRI, the interpretation of the exams and their clinical correlation.
If, in the past, the main concern of the global consensus was about the appropriateness of MRI as a first-line technique, nowadays its role in guiding the management of prostate cancer has been fully established, given its impact on patients’ clinical work-up.
In conclusion, multiple innovations in the field of technology and IT have made uro-radiologists more effective and have increased their contribution to personalised and precision medicine. We encourage all practicing GU radiologists to join and participate in ESUR to help design the future of genitourinary imaging.