Integrating artificial intelligence into the clinical practice of radiology: challenges and recommendations.

Artificial intelligence (AI) has the potential to significantly disrupt the way radiology will be practiced in the near future, but several issues need to be resolved before AI can be widely implemented in daily practice. These include the role of the different stakeholders in the development of AI for imaging, the ethical development and use of AI in healthcare, the appropriate validation of each developed AI algorithm, the development of effective data sharing mechanisms, regulatory hurdles for the clearance of AI algorithms, and the development of AI educational resources for both practicing radiologists and radiology trainees. This paper details these issues and presents possible solutions based on discussions held at the 2019 meeting of the International Society for Strategic Studies in Radiology. KEY POINTS: • Radiologists should be aware of the different types of bias commonly encountered in AI studies, and understand their possible effects. • Methods for effective data sharing to train, validate, and test AI algorithms need to be developed. • It is essential for all radiologists to gain an understanding of the basic principles, potentials, and limits of AI.

Robot-assisted laparoscopic vs open radical cystectomy: comparison of complications and perioperative oncological outcomes in 200 patients.

OBJECTIVE: To compare perioperative morbidity and oncological outcomes of robot-assisted laparoscopic radical cystectomy (RARC) to open RC (ORC) at a single institution. PATIENTS AND METHODS: A retrospective analysis was performed on a consecutive series of patients undergoing RC (100 RARC and 100 ORC) at Wake Forest University with curative intent from 2006 until 2010. Complication data using the Clavien system were collected for 90 days postoperatively. Complications and other perioperative outcomes were compared between patient groups. RESULTS: Patients in both groups had comparable preoperative characteristics. The overall and major complication (Clavien ≥ 3) rates were lower for RARC patients at 35 vs 57% (P = 0.001) and 10 vs 22% (P = 0.019), respectively. There were no significant differences between groups for pathological outcomes, including stage, number of nodes harvested or positive margin rates. CONCLUSION: Our data suggest that patients undergoing RARC have perioperative oncological outcomes comparable with ORC, with fewer overall or major complications. Definitive claims about comparative outcomes with RARC require results from larger, randomised controlled trials.

Testing the performance of risk prediction models to determine progression to referable diabetic retinopathy in an Irish type 2 diabetes cohort.

BACKGROUND /AIMS: To evaluate the performance of existing prediction models to determine risk of progression to referable diabetic retinopathy (RDR) using data from a prospective Irish cohort of people with type 2 diabetes (T2D). METHODS: A cohort of 939 people with T2D followed prospectively was used to test the performance of risk prediction models developed in Gloucester, UK, and Iceland. Observed risk of progression to RDR in the Irish cohort was compared with that derived from each of the prediction models evaluated. Receiver operating characteristic curves assessed models' performance. RESULTS: The cohort was followed for a total of 2929 person years during which 2906 screening episodes occurred. Among 939 individuals followed, there were 40 referrals (4%) for diabetic maculopathy, pre-proliferative DR and proliferative DR. The original Gloucester model, which includes results of two consecutive retinal screenings; a model incorporating, in addition, systemic biomarkers (HbA1c and serum cholesterol); and a model including results of one retinopathy screening, HbA1c, total cholesterol and duration of diabetes, had acceptable discriminatory power (area under the curve (AUC) of 0.69, 0.76 and 0.77, respectively). The Icelandic model, which combined retinopathy grading, duration and type of diabetes, HbA1c and systolic blood pressure, performed very similarly (AUC of 0.74). CONCLUSION: In an Irish cohort of people with T2D, the prediction models tested had an acceptable performance identifying those at risk of progression to RDR. These risk models would be useful in establishing more personalised screening intervals for people with T2D.

Mixed incontinence: does preoperative urodynamic detrusor overactivity affect postoperative quality of life after pubovaginal sling?

OBJECTIVE: Our purpose was to determine if women with mixed urinary incontinence (MUI) and urodynamic detrusor overactivity (DO) have less improvement in urinary symptoms after pubovaginal sling surgery (PVS), compared to MUI without DO. MATERIALS AND METHODS: Women with preoperative MUI symptoms prior to PVS were identified through retrospective review. DO was defined as a symptomatic 5 cm H20 detrusor pressure or greater rise during urodynamics. MUI patients with and without DO before PVS were divided into Groups A and B, respectively. All patients had returned a completed Urogenital Distress Inventory 6 (UDI-6) questionnaire and a 3-day diary of pad usage before surgery and at each postoperative visit. Study endpoints included change in total UDI-6 score, and change in number of pad use/day after PVS. RESULTS: 73 patients were identified, 31 in Group A and 42 in Group B. Mean follow-up after PVS was 15 and 16 months, respectively (p = 0.59). Preoperative total UDI-6 scores were 11.8 and 12.7 (p = 0.30) for Group A and B. Mean changes in total UDI-6 after PVS were - 8.0 and - 10.2 (p = 0.030), respectively. After PVS, both groups reported similar mean reduction in pad/day usage from preoperative baseline (-2.57 vs. --2.49, p = 0.83). There were no differences between the groups when comparing demographic, urodynamic, or operative data. CONCLUSION: MUI patients had improved continence and quality of life after PVS. However, MUI patients with DO had less improvement in UDI-6 scores after PVS, despite a similar reduction to pad use/day.

Facilitated percutaneous coronary intervention versus primary percutaneous coronary intervention: design and rationale of the Facilitated Intervention with Enhanced Reperfusion Speed to Stop Events (FINESSE) trial.

BACKGROUND: Percutaneous coronary intervention (PCI) has emerged as the strategy of choice in reestablishing effective flow in occluded infarct-related arteries in patients with acute myocardial infarction (MI) if it can be administered in a timely fashion. Patients who enter the catheterization laboratory with Thrombolysis In Myocardial Infarction (TIMI) grade 3 blood flow in the infarct-related vessel have better clinical outcomes than patients presenting with impaired flow. We hypothesize that a strategy of early pharmacologic reperfusion therapy with abciximab alone or in conjunction with reduced-dose reteplase, followed by PCI will improve the outcome of patients eligible for primary PCI. STUDY DESIGN: The Facilitated Intervention with Enhanced Reperfusion Speed to Stop Events (FINESSE) study is a 3000-patient, prospective, multicenter, randomized, double-blind, placebo-controlled trial. The study is designed to compare the efficacy and safety of early administration of reduced-dose reteplase and abciximab combination therapy or abciximab alone followed by PCI with abciximab alone administered just before PCI for acute MI. Patients will be randomized to one of these 2 facilitated PCI treatments or primary PCI in a 1:1:1 fashion. The primary efficacy end point of FINESSE is the composite of all-cause mortality or post-MI complications within 90 days of randomization. The primary safety outcome assessment will be Thrombolysis In Myocardial Infarction (TIMI) major bleeding. CONCLUSIONS: The FINESSE study will answer important questions regarding the efficacy and safety of "upstream" medical therapy followed by planned intervention for patients with ST-elevation MI, potentially expanding the population eligible for a primary PCI approach. This study will also provide insight as to which facilitated regimen (reteplase/abciximab combination therapy or abciximab monotherapy) provides the best balance of efficacy and safety.

Retropubic surgery for stress urinary incontinence.

More than 150 surgical techniques have been proposed in the literature for treating stress urinary incontinence. Many of the original published approaches were vaginal, but through the years the literature has expanded to include needle suspension, pubovaginal slings, and retropubic procedures. In this chapter, we focus on retropubic approaches for the treatment of SUI and discuss the physiology, indications, technical details and the complications of these procedures.

Evaluation of the acute coronary syndrome safety profile of dabigatran etexilate in patients undergoing major orthopedic surgery: findings from four Phase 3 trials.

INTRODUCTION: Several anticoagulants have been associated with a 'rebound effect' that potentially increases the risk of thrombosis and cardiovascular events following discontinuation. Four Phase 3 trials of dabigatran etexilate in major orthopedic surgery incorporated measures to assess the risk of acute coronary syndrome (ACS) events during and after treatment. MATERIALS AND METHODS: Patients in RE-MOBILIZE®, RE-MODEL™, RE-NOVATE®, and RENOVATE® II were randomized to dabigatran etexilate (150 mg or 220mg once daily) or enoxaparin for 6-35 days, and followed for up to 90 days. ACS data were tabulated from investigator-reported serious adverse events using ACS-specific Medical Dictionary for Regulatory Authorities (MedDRA) lower-level terms. To ensure that all ACS events were identified in the initial three studies, RE-MOBILIZE®, RE-MODEL™, and RE-NOVATE®, a broader list of MedDRA terms was prespecified that would trigger treatment-blinded adjudication. RESULTS: When pooling the four trials, patients receiving dabigatran etexilate 220 mg had the fewest treatment-emergent, investigator-reported ACS events (6 [0.16%] vs 14 [0.51%] for dabigatran 150 mg and 13 [0.35%] for enoxaparin). Corresponding post-treatment rates were 2 (0.06%), 1 (0.04%), and 4 (0.11%). Similarly, treatment-emergent centrally adjudicated definite or likely ACS events in the first three trials were fewer in patients on dabigatran 220 mg (16 [0.60%]) than dabigatran 150 mg (26 [0.95%]) and enoxaparin (20 [0.74%]). The corresponding numbers post treatment were 2, 2, and 7. None of these between-group differences were statistically significant. CONCLUSION: No increased ACS signal was detected with dabigatran etexilate compared with enoxaparin during or after treatment.

Is robot-assisted radical cystectomy justified in the elderly? A comparison of robotic versus open radical cystectomy for bladder cancer in elderly ≥75 years old.

BACKGROUND AND PURPOSE: Open radical cystectomy (ORC) or minimally invasive radical cystectomy with pelvic lymph node (LN) dissection carries significant morbidity to the elderly because they often have several medical comorbidities that make a surgical approach more challenging. The objective of this study is to compare robot-assisted radical cystectomy (RARC) and ORC in elderly patients. PATIENTS AND METHODS: A prospective bladder cancer cystectomy database was queried to identify all patients age ≥75 years. A total of 20 patients were identified for each of the RARC and ORC cohorts. A retrospective analysis was performed on these 40 patients undergoing radical cystectomy for curative intent. RESULTS: Patients in both groups had comparable preoperative characteristics and demographics. Patients had significant medical comorbidities with 80% in each cohort having American Society of anesthesiologists classification of 3 and 50% having had previous abdominal surgery. Complete median operative times for RARC was 461 (interquartile range [IQR] 331, 554) vs 370 minutes for ORC (IQR 294, 460) (P=0.056); however, median blood loss for RARC was 275 mL (IQR 150, 450) vs 600 mL for ORC (IQR 500, 1925). The median hospital stay for RARC was 7 days (IQR 5, 8) vs 14.5 days for ORC (IQR 8, 22) (P<0.001). The major complication (Clavien≥III) rate for RARC was 10% compared with 35% for ORC (P=0.024). There were two positive margins in the ORC group compared with one in the RARC group with median LN yields of 15 nodes (IQR 11, 22) and 17 nodes (IQR 10, 25) (P=0.560) respectively. CONCLUSIONS: In a comparable cohort of elderly patients, RARC can achieve similar perioperative outcomes without compromising pathologic outcomes, with less blood loss and shorter hospital stays. For an experienced robotic team, RARC should be considered in elderly patients because it may offer significant advantage with respect to perioperative morbidity over ORC.

Changing patterns in iatrogenic ureteral injuries.

Advances in laparoscopic and endoscopic surgery over the past 25 years have changed the preferred methods for performing many operations. We previously reported an increase in the number of patients treated for ureteral injury at our institution that paralleled the introduction of minimally invasive techniques. Since that report, more advanced endoscopic procedures have been introduced. We sought to determine whether the latter influenced the number of ureteral injuries managed at our institution. Reported here are the results of our retrospective study, which sought to determine if the rate of treatment of major iatrogenic ureteral injuries has changed.

Does initial learning curve compromise outcomes for robot-assisted radical cystectomy? A critical evaluation of the first 60 cases while establishing a robotics program.

BACKGROUND AND PURPOSE: Robot-assisted laparoscopic radical cystectomy (RARC) with pelvic lymph node dissection (PLND) has gained popularity as a minimally invasive alternative to open radical cystectomy (ORC) for the treatment of patients with bladder cancer. The learning curve (LC) for laparoscopic and robotic surgery can be steep. We aim to evaluate the effect of the initial LC on operative, postoperative, and pathologic outcomes of the first 60 RARC performed at our newly established robotics program. PATIENTS AND METHODS: After obtaining Institutional Review Board approval, we reviewed the clinical and pathologic data from 60 consecutive patients with clinically localized bladder cancer who underwent RARC with PLND from January 2008 to March 2010. The patients were grouped into tertiles and assessed for effect of LC using analysis of variance. RESULTS: Patient demographics and clinical characteristics were similar across tertiles. The mean total operative time trended down from the 1st to 3rd tertile from 525 minutes to 449 minutes, respectively (P=0.059). Mean estimated blood loss was unchanged across tertiles. Complications decreased as the LC progressed from 14 (70%) in the 1st tertile to 6 (30%) in each of the 2nd and 3rd tertiles (P<0.013). The mean total lymph node yield and number of positive margins were unchanged across tertiles. CONCLUSIONS: RARC with PLND can be performed safely at a high-volume newly established robotic surgery program with an experienced team without compromising operative, postoperative, and short-term pathologic outcomes during the LC for surgeons who are experienced in ORC.

Off-label use of medical devices in radiology: regulatory standards and recent developments.

The US Food and Drug Administration (FDA) enjoys a broad legal mandate to ensure that medical products are safe and effective for their intended uses, authority that extends to the medical devices that make modern radiology possible. Under FDA's regulatory framework, medical devices are cleared or approved for specific indications for use that are described in the labeling provided with the devices. This article outlines this framework as well as the regulatory and legal implications of using medical devices in a manner inconsistent with their FDA-cleared or FDA-approved labeling.

Bladder tissue engineering.

The bladder can lose the ability to store and empty effectively as a result of numerous conditions. When conservative methods to maximize patient safety and quality of life fail, surgical reconstruction of the bladder is usually considered. Augmentation cystoplasty can be performed with the use of the small bowel, large bowel, or less often, stomach. An alternative approach, tissue engineering, identifies the body's own potential for regeneration and supports this propensity with appropriate raw materials and growth factors so that the body's original structure and function may be restored. Tissue engineering can involve the use of a scaffold or matrix alone or of cell-seeded matrices. Harvesting cells and culturing them has become an important tool in tissue engineering. Multiple possibilities for sources of cells have been investigated, including stem cells and differentiated cells from organs other than the bladder; however, to date, autologous bladder cells remain the gold standard for culture and seeding.

Minimizing the cost of surgical correction of stress urinary incontinence and prolapse.

OBJECTIVES: To compare the operative costs associated with the use of incontinence kits, with or without biomaterials, with surgeon-tailored prolene mesh (STPM) in the treatment of stress urinary incontinence (SUI) with or without pelvic organ prolapse. METHODS: All operations for uncomplicated SUI with or without pelvic organ prolapse were reviewed from 2007-2008. Operative billing sheets including operative time, hospital cost, and the insurance billing statement were obtained and reviewed. Surgeon payment was not included in the analysis. Hospital stay was also compared. RESULTS: For patients with SUI alone, there was a significant difference in the hospital cost and the insurance billing statement between STPM and commercial kits (CK). On average, the insurance billing statement for STPM was $2220 less per case as compared with CK. For patients with SUI and anterior compartment prolapse, there was a significant difference in the hospital cost and the insurance billing statement between STPM and CK. On average, the insurance billing statement for STPM was $4770 less per case as compared with CK. For patients with SUI and anterior and posterior compartment prolapse, the difference in hospital cost and insurance billing statement approached statistical significance. The insurance billing statement for STPM on an average was $5600 less per case as compared with CK. There was no significant difference in operative time or hospital stay. CONCLUSIONS: The use of STPM for the treatment of incontinence with or without prolapse is significantly less costly for the hospital and the patient and/or insurance as compared with CK. The use of STPM did not increase operative time or postoperative hospital stay when compared with prefashioned kits.

Self-reported pad use per day reflects patient quality of life after pubovaginal sling surgery.

INTRODUCTION AND HYPOTHESIS: We hypothesized that self-reported pad use per day (PPD) after pubovaginal sling (PVS) correlated with postoperative quality of life (QOL) scores. METHODS: Two hundred fifteen women completed the incontinence impact questionnaire 7 (IIQ-7) and urogenital distress inventory 6 (UDI-6) before PVS and during follow-up. Starting 3 days before a visit, women recorded the number of protective urinary pad changes per day. Analysis of variance and Pearson correlation tests were used to determine if women reporting zero, one, or greater than or equal to two urinary pads per day after PVS had significantly different changes in baseline QOL scores. RESULTS: Over a mean 8.5 months follow-up after PVS, 131, 56, and 28 women reported zero, one, and greater than or equal to two pad changes/day. Each pad group showed progressively less improvement from baseline IIQ-7 and UDI-6 scores after PVS. Change in IIQ-7 and UDI-6 scores negatively correlated with PPD (p < 0.0001). CONCLUSIONS: Self-reported PPD after PVS reflects patient perception of urinary-specific QOL.

Satisfying the requirements of the Food and Drug Administration and Centers for Medicare and Medicare Services.

Clinical studies of new imaging technologies conducted for regulatory purposes have traditionally focused on satisfying the standards of the US Food and Drug Administration (FDA) for marketing clearance or approval of a new product. However, given the increased scrutiny that is being directed at diagnostic imaging by third-party payers, obtaining clinical data that can support Medicare coverage of a new imaging technology is increasingly important to ensure the maximum diffusion of that technology. This article describes the regulatory requirements of both the FDA and Centers for Medicare and Medicare Services and explains how sponsors may approach clinical studies to secure both FDA clearance or approval and Centers for Medicare and Medicare Services coverage and payment.