Retrospective Study of the Impact of Fellowship Training on Two Quality and Safety Measures in Uterine Artery Embolization Sam Stuart, John R. Mayo, Alden Ling, Michael Schulzer, Darren Klass, Mark A. Power, Benjamin J. Roberton, J. M. Wan, David M. Liu, MD
Rationale and Objectives: To measure the impact of 1-year interventional fellowship training on fluoroscopic time and contrast media utilization in uterine artery embolization (UAE). Materials and Methods: Retrospective single institution analysis of 323 consecutive UAEs performed by 12 interventional fellows using a standardized protocol. Fluoroscopy time and contrast media volume were recorded for each patient and correlated with stage of fellowship training. Preprocedure uterine volume (using MRI or ultrasound) was used as a measure of procedural complexity. Regression analysis was conducted per trainee factoring in duration of training, procedure number, supervising radiologist, uterine volume, and outcome variables of fluoroscopy time and contrast media volume. Results: Median number of patients treated per trainee was 27 (range, 16-43) with mean fluoroscopic time 24.5 minutes (range, 4-90 min) and mean contrast volume 190 mL (range, 50-320 mL). Increasing uterine volume had no significant effect (P > .05) on fluoroscopic time but significantly increased (P < .001) contrast media volume. Significant training effect was identified with decrease in fluoroscopic time (P < .001) and decrease in contrast volume (P ¼ .02) over training. Over the course of a 1-year fellowship, these summed to a decrease of 12 minutes in UAE fluoroscopy time and 17 mL less contrast. Conclusion: A significant (P < .05) training effect that is clinically relevant was demonstrated over the course of a yearlong interventional radiology fellowship program in performance of a standardized protocol for UAE. This data supports fellowship training as a basis for UAE credentialing and privileging. Key Words: Credentialing, training, interventional, radiation, uterine artery embolization J Am Coll Radiol 2014;11:471-476. Copyright © 2014 American College of Radiology
INTRODUCTION
Since its description in 1995 [1], uterine artery embolization (UAE) has become an accepted alternative to medical and surgical treatment of uterine leiomyomata (uterine fibroids) [1,2]. A number of large multinational, multi-institutional, prospective, and retrospective studies have shown efficacy and safety for improving clinical outcome variables of pain, bleeding, and symptoms relating to the mass effect of the fibroids in approximately 85% of patients [3-8]. UAE is diagnostically and technically challenging because of difficulties in recognizing the abnormal contrast enhancement pattern of uterine leiomyomata on pelvic angiography and, when identified, safely and completely occluding the abnormal Department of Radiology, Vancouver General Hospital, Vancouver, Canada. Corresponding author and reprints: Dr David M. Liu, MD, Vancouver General Hospital, Department of Radiology, 855 West 12th Ave, Vancouver, BC, V5Z 1M9, Canada; e-mail: [email protected]. There are no relevant conflicts of interest to disclose. ª 2014 American College of Radiology 1546-1440/14/$36.00 http://dx.doi.org/10.1016/j.jacr.2013.09.020
vascular bed. In our experience, 2 clinically relevant easily measured metrics—fluoroscopy time and contrast media volume—can be used as safety metrics that are related to the diagnostic and technical skill level of the operator. Reduction in these 2 measures has been associated with a reduction in procedural risk [9]. We conducted this retrospective study to determine if a training effect could be measured in these 2 metrics within a standardized UAE procedure during a 1-year interventional radiology (IR) fellowship program. MATERIALS AND METHODS
We performed a retrospective (July 2009 to July 2012) single-center study in 323 consecutive women (median age, 45; range, 26-56) treated with UAE for symptomatic uterine leiomyomas. The UAE performance of the 12 clinical IR trainees under the supervision of 7 experienced staff radiologists was the focus of the 471
472 Journal of the American College of Radiology/Vol. 11 No. 5 May 2014
study. We obtained expedited institutional research board ethical approval for this study. Pre-embolization Evaluation
All patients received formal UAE preprocedural consultation by both the referring gynecologist, as well as the IR service, and were deemed appropriate candidates. Standard hematologic panels and coagulation profiles were documented before the procedure. Baseline uterine volume determination was made using MRI (314 of 323, 97%) or ultrasound (9 of 323, 3%). Maximal anterior-posterior, right-left, and cephaladcaudad measurements of the uterus were performed and converted to a volume using the equation for a prolate ellipsoid (d1 d2 d3 .5233) [10]. Clinical Procedural Management
Informed written consent was obtained in all cases. Patients were admitted on the morning of the procedure and brought to the IR suite. Routine intravenous antibiotics were administered in all patients (1 g Cefazolin (Ancef; Smith-Kline Beecham Pharm, Oakville, Ontario or in penicillin allergic patients Clindamycin 300 mg IV; Sando Canada Inc. Qc, Canada). A Foley bladder catheter was inserted by a nurse before treatment. The procedure was performed using conscious/moderate sedation and analgesia with use of midazolam (Versed; Hoffman-LaRoche) and fentanyl narcotic (Sublimaze; Abbott Laboratories, St Laurent, Quebec, Canada) with morphine loaded postprocedural patient-controlled analgesia pump and on-demand antiemetic. All procedures were performed as short-stay procedures with patients initially recovering for 1 to 2 hours in the IR recovery area, then being transferred to the surgical short-stay ward and overnight admission. Nonionic contrast media was used in all cases (Optiray 320, Tyco Healthcare, Montreal, Canada). At discharge, patients were given prescriptions for oral diclofenac (Apotex Inc. Ontario, Canada) and oxycodone with acetaminophen (Percocet; Du Pont Pharma, Mississuaga, Ontario, Canada). Follow-up was performed by the referring gynecologist at 6 weeks and all patients were seen at 3 months for clinical consultation and pelvic ultrasound examination. Embolization Technique
Standard ultrasound-guided Seldinger technique via the right common femoral artery using a 4Fr or 5Fr C1 or C2 catheter and 5F sheath was conducted in all cases. Coaxial microcatheters were not used routinely but reserved for cases in which the operator deemed them necessary. Embolization was performed from a stable and safe position, ideally within the transverse portion in the uterine artery. Polyvinyl alcohol particles, 355 mm to 500 mm, packaged as 1 cc per vial (Contour, Target Therapeutics, Boston Scientific Corporation, Mississauga, Ontario, Canada) was the only embolic agent used. The embolization end point was defined as stasis
of antegrade flow in the uterine artery for at least 5 cardiac cycles. A routine postembolization flush aortogram was performed in all cases. Protocol was maintained for all patients and all supervising staff radiologists. The primary operator for each procedure was 1 of 12 IR fellows each undergoing a 1-year dedicated IR fellowship. All procedures were performed under the direct supervision of 1 of 7 senior staff IR. The fellows were the primary operator for all of the UAE procedures with the supervising physician available in a system of graduated responsibility. Radiation exposure was minimized by the use of pulsed fluoroscopy, collimation, and avoidance of magnification and oblique imaging planes [11]. The retrospective nature of this study blinded operators to interaction with either outcome metric. Statistical Analysis
Analysis was performed on an intention-to-treat basis. Results are expressed as mean, median, and standard error to detail variation in patient and technical details and contrast medium volume and fluoroscopy time. Statistical analysis of the results was performed using the MLwiN multilevel model statistical analysis software program (University of Bristol, UK). Regression analysis was conducted factoring in the individual operator, the number of days of fellowship training, the number of procedures performed during the fellowship, the staff radiologist supervising the procedure, overall fluoroscopy time, contrast medium use, and uterine volume. P values were calculated by Z tests based on the estimated coefficient of regression and the estimated standard error of this coefficient. Statistical significance was determined by a P value .05. RESULTS
Three hundred and twenty three patients underwent UAE, with a mean age of 45 (range, 26-56) and mean uterine volume 642 cc (range, 66-4,828). The average number of patients treated per trainee over the 1-year training was 27 (range, 16-43). Total fluoroscopic time and contrast volume were recorded in 322 of 323 (99.6%) and 321 of 323 (99.3%) patients respectively. The mean total fluoroscopic time was 24.6 minutes per study (range, 3.6-90 minutes) for all fellows. The mean volume of contrast media per UAE was 190 mL, (range, 50-320 mL). Technical Success
Technical success was achieved in 309 of the 323 patients. Failure of bilateral embolization occurred in 14 patients (4%); these patients had unilateral embolization performed. In the cases of failed bilateral embolization, anatomical variation precluding bilateral embolization was the cause in 11 patients. In 9 patients, 1 uterine artery was unilaterally hypoplastic, absent, or too small or tortuous
Stuart et al/Training in Uterine Artery Embolization 473
Fig 1. Significant relationship between the fluoroscopic time and experience of operator. Fluoroscopic time decreases by 1 minute per month of fellowship (standard error 0.44, P < .001).
to catheterize. One uterine artery was providing significant ovarian supply (utero-ovarian anastamoses) in 2 patients and 3 patients had refractory procedurerelated arterial spasm precluding catheterization and embolization. Only 2 patients had repeat procedures— both patients had spasm in 1 of the uterine arteries that prevented successful bilateral embolization at the first attempt. These 2 patients both had unilateral embolization performed during the first procedure. At the second procedure, the opposite uterine artery was successfully embolized. In the remaining 12 patients no further UAE procedure was performed. Effects of Operator Experience
The estimated coefficient of regression from statistical analysis reveals a highly statistically significant decrease in fluoroscopic time of 56 seconds per month (P < .001, 322 of 323 cases in use) in the setting of no statistical variation in the uterine volume on a monthto-month basis, as shown in Figure 1. Furthermore, uterine volume did not have a statistically significant effect on fluoroscopic time (P ¼ .18) with no significant correlation between the number of days of fellowship completed and the uterine volume (P ¼ .35) from hierarchical random effects analysis. Increasing uterine volume had a statistically significant effect on contrast medium volume with contrast medium volume increasing by .02 mL per cubic centimeter of uterus (P < .001). A significant decrease in contrast medium volume of 0.3 mL per week of fellowship (1.2 mL per month) was observed (P ¼ .05) (321 of 323 cases) in the setting of no statistical variation in the uterine volume on a month-to-month basis. With uterine volume adjusted for analysis, the volume of contrast medium administered trended towards a decrease of 0.35 mL per week, 1.5 mL per month, or 17.2 mL per year (P ¼ .02) as shown in Figure 2. Fluoroscopic time decreased with the number of cases performed with a decrease of 0.5 minutes (30 seconds)
Fig 2. The relationship between the contrast medium volume administered (cc) and experience of the operator (days). The volume of contrast medium administered decreases by 1.5 mL per month (standard error 0.02, P ¼ .02).
per case (P < .0001). With a mean number of cases of 27 per fellow, fluoroscopic time decreased by 13 minutes over the course of an average fellowship of 27 cases as shown in Figure 3. The individual staff radiologist supervising the procedure had no effect on the volume of contrast medium administered. There was no significant difference between the 7 supervising staff radiologists, except that 1 of the 7 staff radiologists resulted in a significantly decreased fluoroscopic time compared with 4 of the others. However, at 11 seconds (P < .05), this difference was not clinically relevant. Effects of Early and Late Experience
First four months versus the last four months of fellowship training. The cases performed in the first 4 months of a fellowship had a mean fluoroscopic time of 30.1 minutes (91 of 91 cases) and mean contrast
Fig 3. The relationship between fluoroscopy time (mins) and the number of UAE procedures a fellow has performed during their fellowship. Fluoroscopic time decreases with the number of cases performed with a decrease of 0.5 minutes per case (standard error 0.084, P < .0001). UAE ¼ uterine artery embolization.
474 Journal of the American College of Radiology/Vol. 11 No. 5 May 2014
Table 1. The relationship between the experience of the operator and fluoroscopy time and contrast medium volume Number of Cases Fluoroscopy Time (mins) Contrast Medium Volume (cc) Total First 4 months of fellowship Final 4 months
323 91 122
First 20 cases After 20 cases
220 87
24.5 30.1 21.6 8.5 min decrease 25.8 19.5 6.3 min decrease
medium volume administered of 197 cc (90 of 91 cases) (Table 1). The cases performed in the final third (4 months) of a fellowship had a mean fluoroscopic time of 21.6 minutes (122 of 122 cases) and mean contrast medium volume administered of 186 cc (121 of 122 cases), a decrease of 39% (8.5 minutes) of fluoroscopic time and 22% (41 cc) of contrast medium administered. There was no significant correlation between the uterine volume and the experience of the operator, nor variation in the mean volume of uteri treated within each quartile. Less than 20 cases versus greater than 20 cases performed. Initial 20 procedures performed by a trainee had a mean fluoroscopic time of 26 minutes (standard deviation 14.5, 95% confidence interval, 23.9-27.7) (220 of 220 cases) and contrast medium volume administered of 192 cc (219 of 220 cases). Those cases that were performed after an operator had performed an initial 20 cases had a mean fluoroscopic time of 19.5 minutes and contrast medium volume of 183 cc, a decrease of 25% (6.5 minutes) of fluoroscopic time and 5% (9 cc) of contrast volume administered. There was no significant correlation between the uterine volume and the number of cases an operator had performed. DISCUSSION
As UAE has evolved into a standard of care, controversy remains regarding qualifications and accreditation for performing these procedures. As recently as 2011, concerns regarding radiation exposure, in the setting of patient advocacy, have been identified by the Society of Interventional Radiology as a priority, including the recommendation of recording patient radiation dose in the medical record for fluoroscopically guided procedures [12]. Minimization of radiation exposure, minimization of risk due to excessive contrast medium use, and also technical expertise are required for safe and efficient UAE [9]. The objectives of this study were to identify whether training in a yearlong IR fellowship had an effect on fluoroscopic time and contrast medium volume administered to the patient when performing UAE. Furthermore, this is the first study to report the effects of training on a number of trainees from a single center with a unified, standardized method of performing the procedure, including the exclusive use of nonspherical embolics. This standardization and the
190 197 156 41 cc decrease 192 183 9cc decrease
high sample size of trainees isolates, to the highest degree, the effect of learning on performance. The reported total mean fluoroscopic time of 24.6 minutes is slightly higher than the mean times of 11.0 to 22.5 minutes reported in 5 previous papers as shown in Table 2 [10,13,14-16]. Longer average fluoroscopic times may be due to the relative inexperience of the trainees studied. Indeed, the mean fluoroscopic time of cases performed during the final 4 months of fellowship (21.6 minutes) much closer parallels that of the other studies. The effects of operator experience and fluoroscopic time have been investigated and the literature suggests that increased operator experience decreases fluoroscopic time when performing UAE [10,14]. These studies are limited to experienced staff level IR rather than trainees. Andrews and Brown report that fluoroscopic time varied inversely with operator experience and operator experience is critical in the reduction of fluoroscopic time [16]. The fellows investigated produced a similar decrease in fluoroscopy time of 28.8 seconds per case (or 16.8 minutes over 35 cases). The Andrews and Brown paper demonstrates that 20 typical cases produces reproducibly short fluoroscopy times and recommends that technical proficiency requires the performance of 20 UAE procedures [14]. In comparison, Pron et al report data from a multicenter study including the comparison of fluoroscopic times between early experience and late experience [18] defined as the first 20 and last 20 cases reported, demonstrating a mean fluoroscopy time in early experience of 21.3 minutes and late experience of 16.2 minutes. The decrease in fluoroscopic time of Table 2. Comparing the reported mean fluoroscopic time recorded in the literature on UAE Mean Fluoroscopy Number Time (mins) Author Year of Cases UBC Nikolic et al Andrews et al Nikolic et al Pron et al Vetter et al UBC final 4 months UBC after 20 cases
2012 2000 2000 2001 2003 2004 2012 2012
UAE ¼ uterine artery embolization.
323 18 35 20 570 31 122 87
24.5 21.9 20.4 11.0 18.9 22.5 21.6 19.5
Stuart et al/Training in Uterine Artery Embolization 475
5.1 minutes, or 24% reduction after 20 cases, was significant (P < .001) [10]. This paper, however, did not apply a standardized technique for performing UAE and was drawn from multiple institutions. We present evidence for a statistically significant decrease in fluoroscopic time with experience concordant with the published literature [10,14] with the learning effect extended to the fellowship level. The learning effect shown continues with a trend of improvement throughout the yearlong fellowship, suggesting that during this year, a threshold was not reached. The trainees continued to learn and perform the procedure more effectively and safely, validating the training standards for competency in UAE at 25 procedures set by the SIR in 2001 [9]. Increased operator experience had a trend toward reduction on the total contrast medium volume administered and increased uterine volume had a significant effect on contrast medium dose but not on fluoroscopic time; suggesting the total contrast medium volume administered is, therefore, partly a reflection of operator experience and partly of total uterine volume. The use of contrast medium to visualize particles during embolization means that as uterine volume increases, at least in part a result of increased fibroid volume, more particles and, hence, more contrast medium will be needed to reach the endpoint of embolization, explaining why contrast medium volume was dependent on uterine volume and operator experience. The high degree of technical success (96%) seen concurs with reported American and European studies and meets the suggested threshold of success suggested by the SIR guidelines of 95% [9]. Inability to bilaterally embolize at the first attempt has been reported in 1.8% to 5% of cases [4,17-20]. The results of 4% from this study are comparable. Spasm has been reported to occur in 26% of cases and preventing embolization in 3% [19-21]. The authors acknowledge the potential limitations of this retrospective study. Fluoroscopy time was used as a surrogate of radiation dose to the patient and the true amount of radiation exposure is unknown. Given the high degree of variability in estimation of true dose, we compromised on fluoroscopy time as a surrogate of radiation dose [22]. Furthermore, the role of the staff radiologist (and variation in techniques) could also be viewed as a confounding factor, as all procedures were performed under the direct supervision of a senior staff radiologist. In training at the investigated institution, all fellows are the primary operator for all UAE procedures with a supervising physician available in a system of graduated responsibility. The lack of statistical significance in contrast volume administered and fluoroscopic time with differing supervising staff radiologists suggest that variability in technique of the supervising staff radiologist was not a significant factor in the results seen.
In conclusion, this study demonstrates that time spent training in an IR fellowship increases operator skill and patient safety as measured by significant (P < .05) declines in fluoroscopic time and contrast medium dose. These results are made more robust secondary to the procedural standardization of UAE in this single-center trial with a large sample size of patients, fellows, and uterine volumes. We have documented that, through supervised training, patient safety can be improved by decreasing fluoroscopic time and decreased contrast media usage. TAKE-HOME POINTS
Minimization of radiation exposure, minimization of risk due to excessive contrast medium use, and also technical expertise are required for safe and efficient UAE. This study demonstrates that IR fellowship training and performing UAE results in significant (P < 0.05) declines in fluoroscopic time and contrast medium dose. These results are made more robust secondary to the procedural standardization of UAE in this singlecenter study with a large sample size of patients, fellows and uterine volumes. REFERENCES 1. Goodwin SC, Vedantham S, McLucas B, Forno AE, Perrella R. Preliminary experience with uterine artery embolization for uterine fibroids. J Vasc Interv Radiol 1997;8:517-26. 2. Spies JB, Scialli AR, Jha RC, et al. Initial results from uterine fibroid embolization for symptomatic leiomyomata. J Vasc Interv Radiol 1999;10:1149-57. 3. Hutchins FL Jr, Worthington-Kirsch R, Berkowitz RP. Selective uterine artery embolization as primary treatment for symptomatic leiomyomata uteri. J Am Assoc Gynecol Laparosc 1999;6:279-84. 4. McLucas B, Adler L, Perrella R. Uterine fibroid embolization: nonsurgical treatment for symptomatic fibroids. J Am Coll Surg 2001;92:95-105. 5. Pron G, Bennett J, Common A, et al. The Ontario Uterine Fibroid Embolization Trial. Part 2. Uterine fibroid reduction and symptom relief after uterine artery embolization for fibroids. Fertil Steril 2003;79:120-7. 6. Ravina JH, Ciraru-vigneron N, Aymard A, et al. Uterine artery embolisation for fibroid disease: results of a 6-year study. Minim Invasive Ther Allied Technol 1999;8:441-7. 7. Spies J, Pelage JP. Uterine artery embolization for leiomyomata. Obstet Gynecol 2001;98:29-34. 8. Walker WJ, Pelage JP. Uterine artery embolisation for symptomatic fibroids: clinical results in 400 women with imaging follow up. BJOG 2002;109:1262-72. 9. Spies J, Niedzwiecki G, Goodwin S. Training standards for physicians performing uterine artery embolization for leiomyomata: consensus statement developed by the Task Force on Uterine Artery Embolization and the standards division of the Society of Cardiovascular & Interventional RadiologyeAugust 2000. J Vasc Interv Radiol 2001;12:19-21. 10. Pron G, Bennett J, Common A. Technical results and effects of operator experience on uterine artery embolization for fibroids: the Ontario Uterine Fibroid Embolization Trial. J Vasc Interv Radiol 2003;14:545-54.
476 Journal of the American College of Radiology/Vol. 11 No. 5 May 2014
11. Nikolic B, Abbara S, Levy E, et al. Influence of radiographic technique and equipment on absorbed ovarian dose associated with uterine artery embolization. J Vasc Interv Radiol 2000;11: 1173-8. 12. Tse G, Spies JB. Radiation exposure and uterine artery embolization: current risks and risk reduction. Tech Vasc Interv Radiol 2010;13: 148-53. 13. Nikolic B, Spies JB, Lundsten MJ, Abbara S. Patient radiation dose associated with uterine artery embolization. Radiology 2000;214: 121-5.
17. Braude P, Reidy J, Nott V, Taylor A, Forman R. Embolization of uterine leiomyomata: current concepts in management. Hum Reprod Update 2000;6:603-8. 18. Hutchins FL Jr, Worthington-Kirsch R, Berkowitz RP. Selective uterine artery embolization as primary treatment for symptomatic leiomyomata uteri. J Am Assoc Gynecol Laparosc 1999;6:279-84. 19. Ravina JH, Ravina J, Aymard A, Bouret J, Ciraru-Vigneron N, Houdart E, et al. Particulate arterial embolization: a new treatment for uterine leiomyomata. La Presse Medicale 1998;27:299-303.
14. Andrews RT, Brown PH. Uterine arterial embolization: factors influencing patient radiation exposure. Radiology 2000;217:713-22.
20. Siskin GP, Stainken BF, Dowling K, Meo P, Ahn J, Dolan EG. Outpatient uterine artery embolization for symptomatic uterine fibroids: experience in 49 patients. J Vasc Interv Radiol 2000;11:305-11.
15. Nikolic B, Spies JB, Campbell L, Walsh SM, Abbara S, Lundsten MJ. Uterine artery embolization: reduced radiation with refined technique. J Vasc Interv Radiol 2001;12:39-44.
21. Pelage JP, Soyer P, LeDref O, et al. Uterine arteries: bilateral catheterization with a single femoral approach and a single 5-F catheter— technical note. Radiology 1999;210:573-5.
16. Vetter S, Schultz FW, Strecker EP, Zoetelief J. Patient radiation exposure in uterine artery embolization of leiomyomata: calculation of organ doses and effective dose. Eur Radiol 2004;14:842-8.
22. Miller DL, Balter S, Dixon RG, et al. Quality improvement guidelines for recording patient radiation dose in the medical record for fluoroscopically guided procedures. J Vasc Interv Radiol 2012;23:11-8.
Rationale and Objectives: To measure the impact of 1-year interventional fellowship training on fluoroscopic time and contrast media utilization in uterine artery embolization (UAE). Materials and Methods: Retrospective single institution analysis of 323 consecutive UAEs performed by 12 interventional fellows using a standardized protocol. Fluoroscopy time and contrast media volume were recorded for each patient and correlated with stage of fellowship training. Preprocedure uterine volume (using MRI or ultrasound) was used as a measure of procedural complexity. Regression analysis was conducted per trainee factoring in duration of training, procedure number, supervising radiologist, uterine volume, and outcome variables of fluoroscopy time and contrast media volume. Results: Median number of patients treated per trainee was 27 (range, 16-43) with mean fluoroscopic time 24.5 minutes (range, 4-90 min) and mean contrast volume 190 mL (range, 50-320 mL). Increasing uterine volume had no significant effect (P > .05) on fluoroscopic time but significantly increased (P < .001) contrast media volume. Significant training effect was identified with decrease in fluoroscopic time (P < .001) and decrease in contrast volume (P ¼ .02) over training. Over the course of a 1-year fellowship, these summed to a decrease of 12 minutes in UAE fluoroscopy time and 17 mL less contrast. Conclusion: A significant (P < .05) training effect that is clinically relevant was demonstrated over the course of a yearlong interventional radiology fellowship program in performance of a standardized protocol for UAE. This data supports fellowship training as a basis for UAE credentialing and privileging. Key Words: Credentialing, training, interventional, radiation, uterine artery embolization J Am Coll Radiol 2014;11:471-476. Copyright © 2014 American College of Radiology
INTRODUCTION
Since its description in 1995 [1], uterine artery embolization (UAE) has become an accepted alternative to medical and surgical treatment of uterine leiomyomata (uterine fibroids) [1,2]. A number of large multinational, multi-institutional, prospective, and retrospective studies have shown efficacy and safety for improving clinical outcome variables of pain, bleeding, and symptoms relating to the mass effect of the fibroids in approximately 85% of patients [3-8]. UAE is diagnostically and technically challenging because of difficulties in recognizing the abnormal contrast enhancement pattern of uterine leiomyomata on pelvic angiography and, when identified, safely and completely occluding the abnormal Department of Radiology, Vancouver General Hospital, Vancouver, Canada. Corresponding author and reprints: Dr David M. Liu, MD, Vancouver General Hospital, Department of Radiology, 855 West 12th Ave, Vancouver, BC, V5Z 1M9, Canada; e-mail: [email protected]. There are no relevant conflicts of interest to disclose. ª 2014 American College of Radiology 1546-1440/14/$36.00 http://dx.doi.org/10.1016/j.jacr.2013.09.020
vascular bed. In our experience, 2 clinically relevant easily measured metrics—fluoroscopy time and contrast media volume—can be used as safety metrics that are related to the diagnostic and technical skill level of the operator. Reduction in these 2 measures has been associated with a reduction in procedural risk [9]. We conducted this retrospective study to determine if a training effect could be measured in these 2 metrics within a standardized UAE procedure during a 1-year interventional radiology (IR) fellowship program. MATERIALS AND METHODS
We performed a retrospective (July 2009 to July 2012) single-center study in 323 consecutive women (median age, 45; range, 26-56) treated with UAE for symptomatic uterine leiomyomas. The UAE performance of the 12 clinical IR trainees under the supervision of 7 experienced staff radiologists was the focus of the 471
472 Journal of the American College of Radiology/Vol. 11 No. 5 May 2014
study. We obtained expedited institutional research board ethical approval for this study. Pre-embolization Evaluation
All patients received formal UAE preprocedural consultation by both the referring gynecologist, as well as the IR service, and were deemed appropriate candidates. Standard hematologic panels and coagulation profiles were documented before the procedure. Baseline uterine volume determination was made using MRI (314 of 323, 97%) or ultrasound (9 of 323, 3%). Maximal anterior-posterior, right-left, and cephaladcaudad measurements of the uterus were performed and converted to a volume using the equation for a prolate ellipsoid (d1 d2 d3 .5233) [10]. Clinical Procedural Management
Informed written consent was obtained in all cases. Patients were admitted on the morning of the procedure and brought to the IR suite. Routine intravenous antibiotics were administered in all patients (1 g Cefazolin (Ancef; Smith-Kline Beecham Pharm, Oakville, Ontario or in penicillin allergic patients Clindamycin 300 mg IV; Sando Canada Inc. Qc, Canada). A Foley bladder catheter was inserted by a nurse before treatment. The procedure was performed using conscious/moderate sedation and analgesia with use of midazolam (Versed; Hoffman-LaRoche) and fentanyl narcotic (Sublimaze; Abbott Laboratories, St Laurent, Quebec, Canada) with morphine loaded postprocedural patient-controlled analgesia pump and on-demand antiemetic. All procedures were performed as short-stay procedures with patients initially recovering for 1 to 2 hours in the IR recovery area, then being transferred to the surgical short-stay ward and overnight admission. Nonionic contrast media was used in all cases (Optiray 320, Tyco Healthcare, Montreal, Canada). At discharge, patients were given prescriptions for oral diclofenac (Apotex Inc. Ontario, Canada) and oxycodone with acetaminophen (Percocet; Du Pont Pharma, Mississuaga, Ontario, Canada). Follow-up was performed by the referring gynecologist at 6 weeks and all patients were seen at 3 months for clinical consultation and pelvic ultrasound examination. Embolization Technique
Standard ultrasound-guided Seldinger technique via the right common femoral artery using a 4Fr or 5Fr C1 or C2 catheter and 5F sheath was conducted in all cases. Coaxial microcatheters were not used routinely but reserved for cases in which the operator deemed them necessary. Embolization was performed from a stable and safe position, ideally within the transverse portion in the uterine artery. Polyvinyl alcohol particles, 355 mm to 500 mm, packaged as 1 cc per vial (Contour, Target Therapeutics, Boston Scientific Corporation, Mississauga, Ontario, Canada) was the only embolic agent used. The embolization end point was defined as stasis
of antegrade flow in the uterine artery for at least 5 cardiac cycles. A routine postembolization flush aortogram was performed in all cases. Protocol was maintained for all patients and all supervising staff radiologists. The primary operator for each procedure was 1 of 12 IR fellows each undergoing a 1-year dedicated IR fellowship. All procedures were performed under the direct supervision of 1 of 7 senior staff IR. The fellows were the primary operator for all of the UAE procedures with the supervising physician available in a system of graduated responsibility. Radiation exposure was minimized by the use of pulsed fluoroscopy, collimation, and avoidance of magnification and oblique imaging planes [11]. The retrospective nature of this study blinded operators to interaction with either outcome metric. Statistical Analysis
Analysis was performed on an intention-to-treat basis. Results are expressed as mean, median, and standard error to detail variation in patient and technical details and contrast medium volume and fluoroscopy time. Statistical analysis of the results was performed using the MLwiN multilevel model statistical analysis software program (University of Bristol, UK). Regression analysis was conducted factoring in the individual operator, the number of days of fellowship training, the number of procedures performed during the fellowship, the staff radiologist supervising the procedure, overall fluoroscopy time, contrast medium use, and uterine volume. P values were calculated by Z tests based on the estimated coefficient of regression and the estimated standard error of this coefficient. Statistical significance was determined by a P value .05. RESULTS
Three hundred and twenty three patients underwent UAE, with a mean age of 45 (range, 26-56) and mean uterine volume 642 cc (range, 66-4,828). The average number of patients treated per trainee over the 1-year training was 27 (range, 16-43). Total fluoroscopic time and contrast volume were recorded in 322 of 323 (99.6%) and 321 of 323 (99.3%) patients respectively. The mean total fluoroscopic time was 24.6 minutes per study (range, 3.6-90 minutes) for all fellows. The mean volume of contrast media per UAE was 190 mL, (range, 50-320 mL). Technical Success
Technical success was achieved in 309 of the 323 patients. Failure of bilateral embolization occurred in 14 patients (4%); these patients had unilateral embolization performed. In the cases of failed bilateral embolization, anatomical variation precluding bilateral embolization was the cause in 11 patients. In 9 patients, 1 uterine artery was unilaterally hypoplastic, absent, or too small or tortuous
Stuart et al/Training in Uterine Artery Embolization 473
Fig 1. Significant relationship between the fluoroscopic time and experience of operator. Fluoroscopic time decreases by 1 minute per month of fellowship (standard error 0.44, P < .001).
to catheterize. One uterine artery was providing significant ovarian supply (utero-ovarian anastamoses) in 2 patients and 3 patients had refractory procedurerelated arterial spasm precluding catheterization and embolization. Only 2 patients had repeat procedures— both patients had spasm in 1 of the uterine arteries that prevented successful bilateral embolization at the first attempt. These 2 patients both had unilateral embolization performed during the first procedure. At the second procedure, the opposite uterine artery was successfully embolized. In the remaining 12 patients no further UAE procedure was performed. Effects of Operator Experience
The estimated coefficient of regression from statistical analysis reveals a highly statistically significant decrease in fluoroscopic time of 56 seconds per month (P < .001, 322 of 323 cases in use) in the setting of no statistical variation in the uterine volume on a monthto-month basis, as shown in Figure 1. Furthermore, uterine volume did not have a statistically significant effect on fluoroscopic time (P ¼ .18) with no significant correlation between the number of days of fellowship completed and the uterine volume (P ¼ .35) from hierarchical random effects analysis. Increasing uterine volume had a statistically significant effect on contrast medium volume with contrast medium volume increasing by .02 mL per cubic centimeter of uterus (P < .001). A significant decrease in contrast medium volume of 0.3 mL per week of fellowship (1.2 mL per month) was observed (P ¼ .05) (321 of 323 cases) in the setting of no statistical variation in the uterine volume on a month-to-month basis. With uterine volume adjusted for analysis, the volume of contrast medium administered trended towards a decrease of 0.35 mL per week, 1.5 mL per month, or 17.2 mL per year (P ¼ .02) as shown in Figure 2. Fluoroscopic time decreased with the number of cases performed with a decrease of 0.5 minutes (30 seconds)
Fig 2. The relationship between the contrast medium volume administered (cc) and experience of the operator (days). The volume of contrast medium administered decreases by 1.5 mL per month (standard error 0.02, P ¼ .02).
per case (P < .0001). With a mean number of cases of 27 per fellow, fluoroscopic time decreased by 13 minutes over the course of an average fellowship of 27 cases as shown in Figure 3. The individual staff radiologist supervising the procedure had no effect on the volume of contrast medium administered. There was no significant difference between the 7 supervising staff radiologists, except that 1 of the 7 staff radiologists resulted in a significantly decreased fluoroscopic time compared with 4 of the others. However, at 11 seconds (P < .05), this difference was not clinically relevant. Effects of Early and Late Experience
First four months versus the last four months of fellowship training. The cases performed in the first 4 months of a fellowship had a mean fluoroscopic time of 30.1 minutes (91 of 91 cases) and mean contrast
Fig 3. The relationship between fluoroscopy time (mins) and the number of UAE procedures a fellow has performed during their fellowship. Fluoroscopic time decreases with the number of cases performed with a decrease of 0.5 minutes per case (standard error 0.084, P < .0001). UAE ¼ uterine artery embolization.
474 Journal of the American College of Radiology/Vol. 11 No. 5 May 2014
Table 1. The relationship between the experience of the operator and fluoroscopy time and contrast medium volume Number of Cases Fluoroscopy Time (mins) Contrast Medium Volume (cc) Total First 4 months of fellowship Final 4 months
323 91 122
First 20 cases After 20 cases
220 87
24.5 30.1 21.6 8.5 min decrease 25.8 19.5 6.3 min decrease
medium volume administered of 197 cc (90 of 91 cases) (Table 1). The cases performed in the final third (4 months) of a fellowship had a mean fluoroscopic time of 21.6 minutes (122 of 122 cases) and mean contrast medium volume administered of 186 cc (121 of 122 cases), a decrease of 39% (8.5 minutes) of fluoroscopic time and 22% (41 cc) of contrast medium administered. There was no significant correlation between the uterine volume and the experience of the operator, nor variation in the mean volume of uteri treated within each quartile. Less than 20 cases versus greater than 20 cases performed. Initial 20 procedures performed by a trainee had a mean fluoroscopic time of 26 minutes (standard deviation 14.5, 95% confidence interval, 23.9-27.7) (220 of 220 cases) and contrast medium volume administered of 192 cc (219 of 220 cases). Those cases that were performed after an operator had performed an initial 20 cases had a mean fluoroscopic time of 19.5 minutes and contrast medium volume of 183 cc, a decrease of 25% (6.5 minutes) of fluoroscopic time and 5% (9 cc) of contrast volume administered. There was no significant correlation between the uterine volume and the number of cases an operator had performed. DISCUSSION
As UAE has evolved into a standard of care, controversy remains regarding qualifications and accreditation for performing these procedures. As recently as 2011, concerns regarding radiation exposure, in the setting of patient advocacy, have been identified by the Society of Interventional Radiology as a priority, including the recommendation of recording patient radiation dose in the medical record for fluoroscopically guided procedures [12]. Minimization of radiation exposure, minimization of risk due to excessive contrast medium use, and also technical expertise are required for safe and efficient UAE [9]. The objectives of this study were to identify whether training in a yearlong IR fellowship had an effect on fluoroscopic time and contrast medium volume administered to the patient when performing UAE. Furthermore, this is the first study to report the effects of training on a number of trainees from a single center with a unified, standardized method of performing the procedure, including the exclusive use of nonspherical embolics. This standardization and the
190 197 156 41 cc decrease 192 183 9cc decrease
high sample size of trainees isolates, to the highest degree, the effect of learning on performance. The reported total mean fluoroscopic time of 24.6 minutes is slightly higher than the mean times of 11.0 to 22.5 minutes reported in 5 previous papers as shown in Table 2 [10,13,14-16]. Longer average fluoroscopic times may be due to the relative inexperience of the trainees studied. Indeed, the mean fluoroscopic time of cases performed during the final 4 months of fellowship (21.6 minutes) much closer parallels that of the other studies. The effects of operator experience and fluoroscopic time have been investigated and the literature suggests that increased operator experience decreases fluoroscopic time when performing UAE [10,14]. These studies are limited to experienced staff level IR rather than trainees. Andrews and Brown report that fluoroscopic time varied inversely with operator experience and operator experience is critical in the reduction of fluoroscopic time [16]. The fellows investigated produced a similar decrease in fluoroscopy time of 28.8 seconds per case (or 16.8 minutes over 35 cases). The Andrews and Brown paper demonstrates that 20 typical cases produces reproducibly short fluoroscopy times and recommends that technical proficiency requires the performance of 20 UAE procedures [14]. In comparison, Pron et al report data from a multicenter study including the comparison of fluoroscopic times between early experience and late experience [18] defined as the first 20 and last 20 cases reported, demonstrating a mean fluoroscopy time in early experience of 21.3 minutes and late experience of 16.2 minutes. The decrease in fluoroscopic time of Table 2. Comparing the reported mean fluoroscopic time recorded in the literature on UAE Mean Fluoroscopy Number Time (mins) Author Year of Cases UBC Nikolic et al Andrews et al Nikolic et al Pron et al Vetter et al UBC final 4 months UBC after 20 cases
2012 2000 2000 2001 2003 2004 2012 2012
UAE ¼ uterine artery embolization.
323 18 35 20 570 31 122 87
24.5 21.9 20.4 11.0 18.9 22.5 21.6 19.5
Stuart et al/Training in Uterine Artery Embolization 475
5.1 minutes, or 24% reduction after 20 cases, was significant (P < .001) [10]. This paper, however, did not apply a standardized technique for performing UAE and was drawn from multiple institutions. We present evidence for a statistically significant decrease in fluoroscopic time with experience concordant with the published literature [10,14] with the learning effect extended to the fellowship level. The learning effect shown continues with a trend of improvement throughout the yearlong fellowship, suggesting that during this year, a threshold was not reached. The trainees continued to learn and perform the procedure more effectively and safely, validating the training standards for competency in UAE at 25 procedures set by the SIR in 2001 [9]. Increased operator experience had a trend toward reduction on the total contrast medium volume administered and increased uterine volume had a significant effect on contrast medium dose but not on fluoroscopic time; suggesting the total contrast medium volume administered is, therefore, partly a reflection of operator experience and partly of total uterine volume. The use of contrast medium to visualize particles during embolization means that as uterine volume increases, at least in part a result of increased fibroid volume, more particles and, hence, more contrast medium will be needed to reach the endpoint of embolization, explaining why contrast medium volume was dependent on uterine volume and operator experience. The high degree of technical success (96%) seen concurs with reported American and European studies and meets the suggested threshold of success suggested by the SIR guidelines of 95% [9]. Inability to bilaterally embolize at the first attempt has been reported in 1.8% to 5% of cases [4,17-20]. The results of 4% from this study are comparable. Spasm has been reported to occur in 26% of cases and preventing embolization in 3% [19-21]. The authors acknowledge the potential limitations of this retrospective study. Fluoroscopy time was used as a surrogate of radiation dose to the patient and the true amount of radiation exposure is unknown. Given the high degree of variability in estimation of true dose, we compromised on fluoroscopy time as a surrogate of radiation dose [22]. Furthermore, the role of the staff radiologist (and variation in techniques) could also be viewed as a confounding factor, as all procedures were performed under the direct supervision of a senior staff radiologist. In training at the investigated institution, all fellows are the primary operator for all UAE procedures with a supervising physician available in a system of graduated responsibility. The lack of statistical significance in contrast volume administered and fluoroscopic time with differing supervising staff radiologists suggest that variability in technique of the supervising staff radiologist was not a significant factor in the results seen.
In conclusion, this study demonstrates that time spent training in an IR fellowship increases operator skill and patient safety as measured by significant (P < .05) declines in fluoroscopic time and contrast medium dose. These results are made more robust secondary to the procedural standardization of UAE in this single-center trial with a large sample size of patients, fellows, and uterine volumes. We have documented that, through supervised training, patient safety can be improved by decreasing fluoroscopic time and decreased contrast media usage. TAKE-HOME POINTS
Minimization of radiation exposure, minimization of risk due to excessive contrast medium use, and also technical expertise are required for safe and efficient UAE. This study demonstrates that IR fellowship training and performing UAE results in significant (P < 0.05) declines in fluoroscopic time and contrast medium dose. These results are made more robust secondary to the procedural standardization of UAE in this singlecenter study with a large sample size of patients, fellows and uterine volumes. REFERENCES 1. Goodwin SC, Vedantham S, McLucas B, Forno AE, Perrella R. Preliminary experience with uterine artery embolization for uterine fibroids. J Vasc Interv Radiol 1997;8:517-26. 2. Spies JB, Scialli AR, Jha RC, et al. Initial results from uterine fibroid embolization for symptomatic leiomyomata. J Vasc Interv Radiol 1999;10:1149-57. 3. Hutchins FL Jr, Worthington-Kirsch R, Berkowitz RP. Selective uterine artery embolization as primary treatment for symptomatic leiomyomata uteri. J Am Assoc Gynecol Laparosc 1999;6:279-84. 4. McLucas B, Adler L, Perrella R. Uterine fibroid embolization: nonsurgical treatment for symptomatic fibroids. J Am Coll Surg 2001;92:95-105. 5. Pron G, Bennett J, Common A, et al. The Ontario Uterine Fibroid Embolization Trial. Part 2. Uterine fibroid reduction and symptom relief after uterine artery embolization for fibroids. Fertil Steril 2003;79:120-7. 6. Ravina JH, Ciraru-vigneron N, Aymard A, et al. Uterine artery embolisation for fibroid disease: results of a 6-year study. Minim Invasive Ther Allied Technol 1999;8:441-7. 7. Spies J, Pelage JP. Uterine artery embolization for leiomyomata. Obstet Gynecol 2001;98:29-34. 8. Walker WJ, Pelage JP. Uterine artery embolisation for symptomatic fibroids: clinical results in 400 women with imaging follow up. BJOG 2002;109:1262-72. 9. Spies J, Niedzwiecki G, Goodwin S. Training standards for physicians performing uterine artery embolization for leiomyomata: consensus statement developed by the Task Force on Uterine Artery Embolization and the standards division of the Society of Cardiovascular & Interventional RadiologyeAugust 2000. J Vasc Interv Radiol 2001;12:19-21. 10. Pron G, Bennett J, Common A. Technical results and effects of operator experience on uterine artery embolization for fibroids: the Ontario Uterine Fibroid Embolization Trial. J Vasc Interv Radiol 2003;14:545-54.
476 Journal of the American College of Radiology/Vol. 11 No. 5 May 2014
11. Nikolic B, Abbara S, Levy E, et al. Influence of radiographic technique and equipment on absorbed ovarian dose associated with uterine artery embolization. J Vasc Interv Radiol 2000;11: 1173-8. 12. Tse G, Spies JB. Radiation exposure and uterine artery embolization: current risks and risk reduction. Tech Vasc Interv Radiol 2010;13: 148-53. 13. Nikolic B, Spies JB, Lundsten MJ, Abbara S. Patient radiation dose associated with uterine artery embolization. Radiology 2000;214: 121-5.
17. Braude P, Reidy J, Nott V, Taylor A, Forman R. Embolization of uterine leiomyomata: current concepts in management. Hum Reprod Update 2000;6:603-8. 18. Hutchins FL Jr, Worthington-Kirsch R, Berkowitz RP. Selective uterine artery embolization as primary treatment for symptomatic leiomyomata uteri. J Am Assoc Gynecol Laparosc 1999;6:279-84. 19. Ravina JH, Ravina J, Aymard A, Bouret J, Ciraru-Vigneron N, Houdart E, et al. Particulate arterial embolization: a new treatment for uterine leiomyomata. La Presse Medicale 1998;27:299-303.
14. Andrews RT, Brown PH. Uterine arterial embolization: factors influencing patient radiation exposure. Radiology 2000;217:713-22.
20. Siskin GP, Stainken BF, Dowling K, Meo P, Ahn J, Dolan EG. Outpatient uterine artery embolization for symptomatic uterine fibroids: experience in 49 patients. J Vasc Interv Radiol 2000;11:305-11.
15. Nikolic B, Spies JB, Campbell L, Walsh SM, Abbara S, Lundsten MJ. Uterine artery embolization: reduced radiation with refined technique. J Vasc Interv Radiol 2001;12:39-44.
21. Pelage JP, Soyer P, LeDref O, et al. Uterine arteries: bilateral catheterization with a single femoral approach and a single 5-F catheter— technical note. Radiology 1999;210:573-5.
16. Vetter S, Schultz FW, Strecker EP, Zoetelief J. Patient radiation exposure in uterine artery embolization of leiomyomata: calculation of organ doses and effective dose. Eur Radiol 2004;14:842-8.
22. Miller DL, Balter S, Dixon RG, et al. Quality improvement guidelines for recording patient radiation dose in the medical record for fluoroscopically guided procedures. J Vasc Interv Radiol 2012;23:11-8.
