Tag Archive for: #BJUI

Posts

Article of the month: Current status of artificial intelligence applications in urology and their potential to influence clinical practice

Every month, the Editor-in-Chief selects an Article of the Month from the current issue of BJUI. The abstract is reproduced below and you can click on the button to read the full article, which is freely available to all readers for at least 30 days from the time of this post.

In addition to the article itself, there is an editorial  and a visual abstract produced by prominent members of the urological community. These are intended to provoke comment and discussion and we invite you to use the comment tools at the bottom of each post to join the conversation. 

If you only have time to read one article this month, it should be this one.

Current status of artificial intelligence applications in urology and their potential to influence clinical practice

Jian Chen*, Daphne Remulla*, Jessica H. Nguyen*, D. Aastha, Yan Liu, Prokar Dasgupta and Andrew J. Hung*

*Catherine & Joseph Aresty Department of Urology, Center for Robotic Simulation & Education, University of Southern California Institute of Urology, Computer Science Department, Viterbi School of Engineering, University of Southern California, Los Angeles, CA, USA, and Division of Transplantation Immunology and Mucosal Biology, Faculty of Life Sciences and Medicine, Kings College London, London, UK

Abstract

Objective

To investigate the applications of artificial intelligence (AI) in diagnosis, treatment and outcome prediction in urologic diseases and evaluate its advantages over traditional models and methods.

Materials and methods

A literature search was performed after PROSPERO registration (CRD42018103701) and in compliance with Preferred Reported Items for Systematic Reviews and Meta‐Analyses (PRISMA) methods. Articles between 1994 and 2018 using the search terms “urology”, “artificial intelligence”, “machine learning” were included and categorized by the application of AI in urology. Review articles, editorial comments, articles with no full‐text access, and non-urologic studies were excluded.

Results

Initial search yielded 231 articles, but after excluding duplicates and following full‐text review and examination of article references, only 111 articles were included in the final analysis. AI applications in urology include: utilizing radiomic imaging or ultrasonic echo data to improve or automate cancer detection or outcome prediction, utilizing digitized tissue specimen images to automate detection of cancer on pathology slides, and combining patient clinical data, biomarkers, or gene expression to assist disease diagnosis or outcome prediction. Some studies employed AI to plan brachytherapy and radiation treatments while others used video based or robotic automated performance metrics to objectively evaluate surgical skill. Compared to conventional statistical analysis, 71.8% of studies concluded that AI is superior in diagnosis and outcome prediction.

Conclusion

AI has been widely adopted in urology. Compared to conventional statistics AI approaches are more accurate in prediction and more explorative for analyzing large data cohorts. With an increasing library of patient data accessible to clinicians, AI may help facilitate evidence‐based and individualized patient care.

 

Editorial: Machines in urology: a brief odyssey of the future

Artificial intelligence (AI) will bring in a new wave of changes in the medical field, likely altering how we practice medicine. In a timely contribution, Chen et al. [1] outline the current landscape of AI and provide us with a glimpse of the future, in which sophisticated computers and algorithms play a front-and-centre role in the daily hospital routine.

Widespread adoption of electronic medical records (EMRs), an ever-increasing amount of radiographic imaging, and the ubiquity of genome sequencing, among other factors, have created an impossibly large body of medical data. This poses obvious challenges for clinicians to remain abreast of new discoveries, but also presents new opportunities for scientific discovery. AI is the inevitable and much-needed tool with which to harness the ‘big data’ of medicine.

Currently, the most immediate and important application of AI appears to be in the field of diagnostics and radiology. In prostate cancer, for example, machine learning algorithms (MLAs) are not only able to automate radiographic detection of prostate cancer but have also been shown to improve diagnostic accuracy compared to standard clinical scoring schemes. MLAs can use clinicopathological data to predict clinically significant prostate cancer and disease recurrence
with a high degree of accuracy. The same has been shown for other urological malignancies, including urothelial cancer and RCC. Implementation of MLAs will lead to improved accuracy and reproducibility, reducing human bias and variability. We also predict that as natural language processing becomes more sophisticated, the troves of nonstructured data that exist in EMRs will be harnessed to deliver improved and more personalized patient care. Patient data and clinical outcomes can be analysed in short time, drawing from a deep body of knowledge, and leading to rapid insights that can guide medical decision-making.

Current AI technology, however, remains experimental and we are still far from the widespread implementation of AI within clinical medicine. A valid criticism of today’s AI is that it functions in the setting of a ‘black box’; the rules that govern the clinical decision-making of an algorithm are often poorly understood or unknowable. We cannot become operators of machines for which we know not how they work, to do so would be to practice medicine blindly.

Another barrier to incorporating AI into common practice is the level of noise in healthcare data. MLAs will use whatever data that are fed to the algorithm, thus running the risk of producing predicative models that include nonsensical variables gleaned from the noise. This concept is similar to multiple hypothesis-testing, where if you feed enough random information into a model, a pattern might emerge. Furthermore, none of the studies described by Chen et al. have been externally validated on large, representative datasets of diverse patients. MLAs trained on a narrow patient population run the risk of creating predictions that
are not generalizable. This problem has already been popularized within genome analysis, where one study found that 81% of all genome-wide studies were taken from individuals of European ancestry [2]. It is easy to imagine situations where risk score calculators or biomarkers are validated using non-representative datasets, leading to less accurate and even inappropriate treatment decisions for underrepresented patient populations. At best, MLAs that are not validated using stringent principles can lead to erroneous disease models. At worst, they can bias the delivery of healthcare to patients, leading to worse patient outcomes and exacerbation of healthcare disparities.

Chen et al. write of the possibility of AI in urology today. What about the future? Imagine a world in which computers with a robotic interface see patients in clinics, design and carry out complex medical treatment plans, and perform surgery without the aid of a human hand. This future may not be far off [3]. Or, even stranger, consider a world in which generalizable AI exists. Estimates of the dawn of this technology range, however the most optimistic projections put the timeline on the order of 20–30 years. Not far behind could be the ‘singularity’, a moment when technological advancement occurs at such an exponential rate that improbable scientific discoveries happen almost instantaneously, setting off a feed-forward cycle leading to an inconceivable superintelligence.

The future is, of course, hard to predict. Nevertheless, AI and the ensuing technology will certainly transform the practice of urology, albeit not without significant challenges and growing pains along the way. The urologist of the future may look very different indeed.

by Stephen W. Reese, Emily Ji, Aliya Sahraoui and Quoc-Dien Trinh

 

References

  1. Chen J, Remulla D, Nguyen JH et al. Current status of artificial intelligence applications in Urology and its potential to influence clinical practice. BJU Int 2019; 124: 567–77
  2. Popejoy AB, Fullerton SM. Genomics is failing on diversity. Nature 2016; 538: 161–4
  3. Grace K, Salvatier J, Dafoe A, Zhang B, Evans O. When Will AI Exceed Human Performance? Evidence from AI Experts, 2017

 

Video: Current status of artificial intelligence applications in urology

Current status of artificial intelligence applications in urology and their potential to influence clinical practice

Abstract

Objective

To investigate the applications of artificial intelligence (AI) in diagnosis, treatment and outcome prediction in urologic diseases and evaluate its advantages over traditional models and methods.

Materials and methods

A literature search was performed after PROSPERO registration (CRD42018103701) and in compliance with Preferred Reported Items for Systematic Reviews and Meta‐Analyses (PRISMA) methods. Articles between 1994 and 2018 using the search terms “urology”, “artificial intelligence”, “machine learning” were included and categorized by the application of AI in urology. Review articles, editorial comments, articles with no full‐text access, and non-urologic studies were excluded.

Results

Initial search yielded 231 articles, but after excluding duplicates and following full‐text review and examination of article references, only 111 articles were included in the final analysis. AI applications in urology include: utilizing radiomic imaging or ultrasonic echo data to improve or automate cancer detection or outcome prediction, utilizing digitized tissue specimen images to automate detection of cancer on pathology slides, and combining patient clinical data, biomarkers, or gene expression to assist disease diagnosis or outcome prediction. Some studies employed AI to plan brachytherapy and radiation treatments while others used video based or robotic automated performance metrics to objectively evaluate surgical skill. Compared to conventional statistical analysis, 71.8% of studies concluded that AI is superior in diagnosis and outcome prediction.

Conclusion

AI has been widely adopted in urology. Compared to conventional statistics AI approaches are more accurate in prediction and more explorative for analyzing large data cohorts. With an increasing library of patient data accessible to clinicians, AI may help facilitate evidence‐based and individualized patient care.

What’s the diagnosis?

These images from Fraisse et al (BJUI 2019) who compare outcomes of treating T1 renal tumours with partial nephrectomy and percutaneous cryoablation show post treatment at 1 day and 1, 3 and 6 months (from left to right).

No such quiz/survey/poll

Article of the week: Targeted deep sequencing of urothelial bladder cancers and associated urinary DNA: a 23‐gene panel with utility for non‐invasive diagnosis and risk stratification

Every week, the Editor-in-Chief selects an Article of the Week from the current issue of BJUI. The abstract is reproduced below and you can click on the button to read the full article, which is freely available to all readers for at least 30 days from the time of this post.

In addition to the article itself, there is an editorial written by a prominent member of the urological community and a video prepared by the authors. These are intended to provoke comment and discussion and we invite you to use the comment tools at the bottom of each post to join the conversation. 

If you only have time to read one article this week, it should be this one.

Targeted deep sequencing of urothelial bladder cancers and associated urinary DNA: a 23‐gene panel with utility for non‐invasive diagnosis and risk stratification

Douglas G. Ward*, Naheema S. Gordon*, Rebecca H. Boucher*, Sarah J. Pirrie*, Laura Baxter, Sascha Ott, Lee Silcock, Celina M. Whalley*, Joanne D. Stockton*, Andrew D. Beggs*, Mike Griffiths§, Ben Abbotts*, Hanieh Ijakipour*, Fathimath N.Latheef*, Robert A. Robinson*, Andrew J. White*, Nicholas D. James*, Maurice P.Zeegers, K. K. Cheng** and Richard T. Bryan*

 

*Institute of Cancer and Genomic Sciences, University of Birmingham, Birmingham, Department of Computer Science, University of Warwick, Coventry, Nonacus Limited, Birmingham Research Park, §West Midlands Regional Genetics Laboratory, Birmingham Women’s and Children’s NHS Foundation Trust, Birmingham, UK, NUTRIM School for Nutrition and Translational Research in Metabolism and CAPHRI Care and Public Health Research Institute, Maastricht University, Maastricht, The Netherlands and **Institute of Applied Health Research, University of Birmingham, Birmingham, UK

Abstract

Objectives

To develop a focused panel of somatic mutations (SMs) present in the majority of urothelial bladder cancers (UBCs), to investigate the diagnostic and prognostic utility of this panel, and to compare the identification of SMs in urinary cell‐pellet (cp) DNA and cell‐free (cf) DNA as part of the development of a non‐invasive clinical assay.

Patients and Methods

A panel of SMs was validated by targeted deep‐sequencing of tumour DNA from 956 patients with UBC. In addition, amplicon and capture‐based targeted sequencing measured mutant allele frequencies (MAFs) of SMs in 314 urine cpDNAs and 153 urine cfDNAs. The association of SMs with grade, stage and clinical outcomes was investigated by univariate and multivariate Cox models. Concordance between SMs detected in tumour tissue and cpDNA and cfDNA was assessed.

 

Results

The panel comprised SMs in 23 genes: TERT (promoter), FGFR3, PIK3CA, TP53, ERCC2, RHOB, ERBB2, HRAS, RXRA, ELF3, CDKN1A, KRAS, KDM6A, AKT1, FBXW7, ERBB3, SF3B1, CTNNB1, BRAF, C3orf70, CREBBP, CDKN2A and NRAS; 93.5–98.3% of UBCs of all grades and stages harboured ≥1 SM (mean: 2.5 SMs/tumour). RAS mutations were associated with better overall survival (P = 0.04). Mutations in RXRA, RHOB and TERT (promoter) were associated with shorter time to recurrence (P < 0.05). MAFs in urinary cfDNA and cpDNA were highly correlated; using a capture‐based approach, >94% of tumour SMs were detected in both cpDNA and cfDNA.

Conclusions

SMs are reliably detected in urinary cpDNA and cfDNA. The technical capability to identify very low MAFs is essential to reliably detect UBC, regardless of the use of cpDNA or cfDNA. This 23‐gene panel shows promise for the non‐invasive diagnosis and risk stratification of UBC.

 

Editorial: Non‐invasive diagnosis and monitoring of urothelial bladder cancer: are we there yet?

In this issue of BJUI, Ward et al. [1] describe the development of DNA‐based urinary biomarkers for urothelial carcinoma (UC). The genomics of UC have been well characterized through interrogation of tumour issues in institutional series (e.g. the Memorial Sloan Kettering Cancer Center [MSKCC] experience), multi‐institutional collaborations (e.g. The Cancer Genome Atlas [TCGA]) and commercial platforms (e.g. the Foundation Medicine experience) [2]. Until recently, these have been largely academic pursuits, with possible impact on prognostication but limited clinical applicability and utility for therapy selection and monitoring of response; however, with the US Food and Drug Administration approval of erdafitinib several weeks ago, patients with advanced UC will routinely receive genomic assessment for FGFR2/3 mutation or fusion, the targets for this therapy [3]. In due time, it is anticipated that multiple other putative targets with associated therapies (e.g. ERBB2, CDKN2A), as well as potential predictive biomarkers, may also warrant testing.

The evolving landscape in advanced UC makes a non‐invasive biomarker particularly attractive. The authors of the present commentary have previously reported results from a series of 369 patients with advanced UC, demonstrating that genomic alterations in ctDNA could be identified in 91% of patients using a commercially available 73-gene panel [4]. More recently, Christensen et al. [5] assessed a cohort of 68 patients receiving neoadjuvant chemotherapy for muscle‐invasive disease, demonstrating 100% sensitivity and 98% specificity for the detection of relapsed disease with a patient‐specific ctDNA assessment (sequenced to a median target coverage of 105 000×) after cystectomy. Impressively, the data also showed that the dynamics of ctDNA appeared to be more useful than pathological downstaging in predicting relapse.

In contrast to these studies, Ward et al. have developed a 23‐gene panel based on frequently expressed genes in a cohort of 916 UC tissue specimens, largely derived from patients with non‐muscle‐invasive disease. Ultimately, with a cohort of 314 patients with DNA derived from a urinary cell pellet, sequencing identified 645 (71.4%) of 903 mutations detected in tumour. Using urinary supernatant, 353 (80.7%) of 437 mutations were detected. These relatively high sensitivities, if they can be interpreted as such, are promising but do not rise to the level of replacing existing strategies for UC detection, staging and monitoring. Notably, another study demonstrated that urinary ctDNA can be detected with high sensitivity and specificity in patients with localized early‐stage bladder cancer and for after‐treatment surveillance, providing the foundation for further studies evaluating the role of ctDNA in non‐invasive detection, genotyping and monitoring [6].

Beyond its use as a diagnostic tool, it is hoped that urinary ctDNA may also find applications in the selection of therapeutics. To this end, Ward et al. identified FGFR3, PIK3CA, ERCC2 and ERBB2 mutations in 45%, 32%, 14% and 7% of patients, respectively. The frequency of FGFR3 alteration decreased with increasing stage and grade, ranging from 72% in pTaG1 disease to just 13% in ≥pT2 disease, consistent with other reports [7]. These results may guide forthcoming studies evaluating FGFR inhibitors in non‐muscle‐invasive, muscle‐invasive and metastatic disease, where studies are ongoing. In reviewing the potential link between genomic alterations and clinical outcomes, perhaps the most curious finding is that between RAS mutations and improved overall survival (P = 0.04), the only such association found in multivariate analysis. These results stand in sharp contrast to reports in lung cancer, colorectal cancer and multiple other tumour types [8]. A closer look at the deleterious nature and functional impact of NRAS and KRAS mutations seen in this series is certainly warranted, along with further external validation in a more homogenous and larger patient population. There is also the potential application of monitoring treatment response by assessing eradication of urinary ctDNA, a hypothesis that is being evaluated in ongoing studies [9].

How will the results of this and other emerging urinary biomarker studies eventually make their way to the clinic? The answer is simple: incorporation of these biomarkers in prospective therapeutic trials. As the bladder cancer investigative community formulates novel trials for non‐muscle‐invasive and muscle‐invasive disease using targeted therapies, an excellent opportunity exists to correlate urinary, blood and tissue‐based biomarkers and to assess their relative predictive capabilities and clinical utility. Furthermore, with clinical surrogate endpoints likely to drive regulatory approval (e.g. landmark complete response rates for non‐muscle‐invasive disease, or pT0N0 rate for muscle‐invasive disease), a validated urinary biomarker could ultimately offer an alternative biological surrogate endpoint [10]. In an era of genomic revolution, prospective validation can help establish the potential clinical utility of promising biomarkers and help realize the dream of ‘precision oncology’.

by Rohit K. Jain, Petros Grivas and Sumanta K. Pal

References

  1. Ward DGGordon NSBoucher RH et al. Targeted deep sequencing of urothelial bladder cancers and associated urinary DNA: a 23‐gene panel with utility for non‐invasive diagnosis and risk stratification. BJU Int 2019
  2. Schiff JPBarata PCYu EYGrivas PPrecision therapy in advanced urothelial cancer. Expert Rev Precis Med Drug Dev 2019481– 93
  3. FDA grants accelerated approval to erdafitinib for metastatic urothelial carcinoma [press release] 2019.
  4. Agarwal NPal SKHahn AW et al. Characterization of metastatic urothelial carcinoma via comprehensive genomic profiling of circulating tumor DNA. Cancer 20181242115– 24
  5. Christensen EBirkenkamp‐Demtroder KSethi H et al. Early detection of metastatic relapse and monitoring of therapeutic efficacy by ultra‐deep sequencing of plasma cell‐free DNA in patients with urothelial bladder carcinoma. J Clin Oncol 2019371547– 57
  6. Dudley JCSchroers‐Martin JLazzareschi DV et al. Detection and surveillance of bladder cancer using urine tumor DNA. Cancer Discov 20199500– 9
  7. Tomlinson DCBaldo OHarnden PKnowles MAFGFR3 protein expression and its relationship to mutation status and prognostic variables in bladder cancer. J Pathol 200721391– 8
  8. Zhuang RLi SLi Q et al. The prognostic value of KRAS mutation by cell‐free DNA in cancer patients: a systematic review and meta‐analysis. PLoS One 201712e0182562
  9. Abbosh PHPlimack ERMolecular and clinical insights into the role and significance of mutated DNA repair genes in bladder cancer. Bladder Cancer 201849– 18
  10. Jarow JPLerner SPKluetz PG et al. Clinical trial design for the development of new therapies for nonmuscle‐invasive bladder cancer: report of a Food and Drug Administration and American Urological Association public workshop. Urology 201483262– 4

 

 

Video: Targeted deep sequencing of urothelial bladder cancers and associated urinary DNA

Targeted deep sequencing of urothelial bladder cancers and associated urinary DNA: a 23‐gene panel with utility for non‐invasive diagnosis and risk stratification

Abstract

Objectives

To develop a focused panel of somatic mutations (SMs) present in the majority of urothelial bladder cancers (UBCs), to investigate the diagnostic and prognostic utility of this panel, and to compare the identification of SMs in urinary cell‐pellet (cp) DNA and cell‐free (cf) DNA as part of the development of a non‐invasive clinical assay.

Patients and Methods

A panel of SMs was validated by targeted deep‐sequencing of tumour DNA from 956 patients with UBC. In addition, amplicon and capture‐based targeted sequencing measured mutant allele frequencies (MAFs) of SMs in 314 urine cpDNAs and 153 urine cfDNAs. The association of SMs with grade, stage and clinical outcomes was investigated by univariate and multivariate Cox models. Concordance between SMs detected in tumour tissue and cpDNA and cfDNA was assessed.

Results

The panel comprised SMs in 23 genes: TERT (promoter), FGFR3, PIK3CA, TP53, ERCC2, RHOB, ERBB2, HRAS, RXRA, ELF3, CDKN1A, KRAS, KDM6A, AKT1, FBXW7, ERBB3, SF3B1, CTNNB1, BRAF, C3orf70, CREBBP, CDKN2A and NRAS; 93.5–98.3% of UBCs of all grades and stages harboured ≥1 SM (mean: 2.5 SMs/tumour). RAS mutations were associated with better overall survival (P = 0.04). Mutations in RXRA, RHOB and TERT (promoter) were associated with shorter time to recurrence (P < 0.05). MAFs in urinary cfDNA and cpDNA were highly correlated; using a capture‐based approach, >94% of tumour SMs were detected in both cpDNA and cfDNA.

Conclusions

SMs are reliably detected in urinary cpDNA and cfDNA. The technical capability to identify very low MAFs is essential to reliably detect UBC, regardless of the use of cpDNA or cfDNA. This 23‐gene panel shows promise for the non‐invasive diagnosis and risk stratification of UBC.

 

Making real change where it is needed! The HSIB investigation into a case of testicular torsion

This week saw the first Health Service Investigation Branch (HSIB) investigation into a urological condition. The HSIB is the health services version of the Air Investigation Branch, which investigate air crashes, and the case that it was investigating was one of testicular loss from torsion.

The investigation followed the best principles of human factors theory and causal analysis. It was not looking to assign blame but instead to constructively implement better process and systems that do not relay solely on one individual, as humans are notoriously fallible. The outcome of any investigation is to make it easier for medical teams and administrators to perform well and to mitigate the risk of errors, in an inherently complex area such as medicine.

Only a small number of HSIB investigations have taken place so far so we are fortunate that a Urology case was chosen. The report concentrated on the community aspects of the testicular pain pathway, and the investigating team had fruitful meetings with NHS 111 that led to changes in the questions and prompts that were asked of callers with testicular pain who dialled in. The Royal College of GPs, as a result of the investigation, has convened a group to review the communication standards between practices running telephone services and emergency departments; and NICE has agreed to improve the on-line guidance on testicular torsion and scrotal pain to make it more accessible to clinicians, patients and their carers.

The fact that this came about after an investigation of a single case shows the power of this investigative process and the rigour with which it was carried out.

I would encourage others who may want to be involved with this type of work. I was lucky enough to be approached by the HSIB to be the subject matter expert (SME) on this case as I have a known interest in both Quality Improvement (QI) and Torsion. Anyone approached to help with investigations of this type should be reassured of the professionalism under which a case is undertaken: no individuals or organisations are named; no fingers are pointed but instead the HSIB are able to open a lot of doors and instigate change by negotiating agreements from departments and institutions that most clinicians involved in QI could only dream of getting.

Maybe we need a few more investigations of this type in Urology; retained stents spring immediately to mind as a strong candidate as the HSIB is also experienced in talking to industry. Wouldn’t retained stents be so much easier to avoid if each stent had an individualised barcode that could be scanned and tracked? The companies making stents could perhaps be encouraged to be more involved in making sure that they were easier to track across the whole of the UK (or the world) so patients wouldn’t have so many problems with stents in the future.  Every component of a jet is tracked in a similar way so why shouldn’t we look for the same standard in Urology Healthcare!

by Tony Tien & James Green

Twitter: @greenxmedical

 

James S. A. Green is a Urological Surgeon, Network Lead for Urology at Barts Health NHS Trust, Quality Improvement Director at Whipps Cross University Hospital and visiting Professor in Health Services Research at Kings College, London. His interest in medical education and improvement started when developing medical support for the British Army and he has published extensively on team-working and improving clinical care. He was SME for the HSIB investigation into a case of delayed testicular torsion.

Mr Tony Tien MRCS is a clinical fellow in Urology at Whipps Cross Hospital and a champion for Quality Improvement.

 

© 2019 BJU International. All Rights Reserved.