Comparing the Quality of the Suture Anastomosis and the Learning Curves Associated with Performing...

CaPLCKH

Tiadb

DSF

R2FCsMiCCM

©P

omparing the Quality of the Suture Anastomosisnd the Learning Curves Associated witherforming Open, Freehand, and Robotic-Assistedaparoscopic Pyeloplasty in a Swine Animal Modelarlo C Passerotti, MD, PhD, Ana Maria AMS Passerotti, MD, Marcos F Dall’Oglio, MD, PhD,atia RM Leite, MD, PhD, Ricardo LV Nunes, MD, Miguel Srougi, MD, Alan B Retik, MD, FACS,iep T Nguyen, MD

BACKGROUND: It is believed that robotic assistance allows for improved suture reapproximation of tissue anddecreases the lengthy learning time that is needed to master laparoscopic suturing. But therehave been no studies directly comparing the efficiency of robotic-assisted laparoscopic surgery(RALS) to freehand laparoscopy (LS) and open surgery (OS). The purpose of this study was tocompare the quality of the suture anastomosis of the ureteropelvic junction (UPJ) using thethree techniques and to evaluate their associated learning curves.

STUDY DESIGN: The operative time for dismembered pyeloplasties performed in 57 pigs by 3 inexperienced and1 experienced surgeon using each of the techniques was measured. The anastomosis was eval-uated for water tightness and patency using antegrade and retrograde urodynamic measure-ments immediately after surgery and 2 weeks postoperatively.The histology of the operated UPJwas also evaluated at 15 days postoperatively.

RESULTS: RALS had a shorter procedural time and less steep learning curve compared with LS. Urody-namic measurements for patency and water tightness of the UPJ were comparable to those inthe OS group. But with experience, both the RALS and LS procedural times and the urody-namic measurements for water tightness and patency of the UPJ approached those of the OSgroup. Histologic evaluation demonstrated that there was less collagen III deposition aroundthe operated UPJ in pigs that underwent RALS compared with LS and OS.

CONCLUSIONS: Among inexperienced surgeons, the efficiency of performing suturing using RALS is operatorindependent, requires less time to learn, and is better than those done by LS technique. (J Am
Coll Surg 2009;208:576–586. © 2009 by the American College of Surgeons)
Rtarvfclcl

h(gnd

he role of robotic assistance in surgery was first popular-zed in the 1980s, with its initial application as navigatorsnd stereotactic localizers in neurosurgery.1 Over the last 2ecades, its role has rapidly expanded, and it is currentlyeing used routinely in various laparoscopic procedures.

isclosure Information: Nothing to disclose.upport for this study was received from CAPES and FAPESP, Brazilianunding Agencies.

eceived November 12, 2008; Revised January 1, 2009; Accepted January 7,009.rom the Robotic Research and Training Center, Department of Urology,hildren’s Hospital Boston, Harvard Medical School, Boston, MA (CC Pas-

erotti, AM Passerotti, Dall’Oglio, Retik, Nguyen) and the Laboratory ofedical Investigation, Department of Urology, University of Sao Paulo Med-

cal School, Sao Paulo, Brazil (Dall’Oglio, Leite, Nunes, Srougi).orrespondence address: Hiep T Nguyen, MD, Department of Urology,hildren’s Hospital Boston, 300 Longwood Ave, Hunnewell-353, Boston,

pA 02115.

5762009 by the American College of Surgeons

ublished by Elsevier Inc.

obotic assistance has played a pivotal role in increasinghe application of minimally invasive surgery in the pedi-tric population.2 It is believed that robotic-assisted lapa-oscopy (RALS), with its 3-dimensional, enhanced (10�)ision and greater degree of rotational movement, allowsor improved suture reapproximation of tissue and de-reases the lengthy learning time that is needed to masteraparoscopic suturing.3,4 Consequently, more complex re-onstructive procedures are now being performedaparoscopically.

Several open and minimally invasive surgical techniquesave been used in the treatment of ureteropelvic junctionUPJ) obstruction. Although open surgery remains theold standard, achieving up to a 97% success rate,5,6 alter-ative techniques have been developed in the hope of re-ucing the morbidity associated with performing an open
yeloplasty while maintaining the high success rate. Re-
ISSN 1072-7515/09/$36.00doi:10.1016/j.jamcollsurg.2009.01.010

csstsdvwtuwrn

MTvFgrdbajwtnTwreeepppfaese

STmmw4e(a7v

gitSfumawstpatm

sufitTfrmTrmspctv

EAuts

577Vol. 208, No. 4, April 2009 Passerotti et al Outcomes in Performing Laparoscopic Pyeloplasty

ently, robotic-assisted laparoscopic pyeloplasty has beenhown to be equally effective in the treatment of UPJ ob-truction, with a 90% to 95% success rate.7-9 The success ofhe surgery is attributed to the improved visualization anduturing technique offered by the robotic assistance. But toate, there has been no study evaluating the utility or ad-antage of performing suturing using RALS comparedith open (OS) and freehand laparoscopic (LS) surgery

echniques. In this study, we evaluated the quality of thereteropelvic anastomosis and the learning curve associatedith performing the pyeloplasty using open, freehand lapa-

oscopic, and robotic-assisted laparoscopic techniques byonexperienced surgeons.

ETHODShis study was approved and performed under the super-ision of our institution’s animal research committee.ifty-seven swine (weighing 30 kg) were distributed into 3roups: open surgery (OS), freehand laparoscopy (LS), andobotic-assisted laparoscopy (RALS). We elected to use leftismembered pyeloplasty as the surgical model to studyecause of the constant length of the anastomosis and thebility to evaluate the quality of the anastomosis using ob-ective urodynamic criteria. The principles of the operationere identical in all three groups. The only differences were

he method of access to the kidney and the suturing tech-ique used to anastomose the ureter to the renal pelvis.hree surgeons performed the pyeloplasty. S1 (AMP)as a nonurologic surgeon who did not have any expe-

ience in LS or RALS. S2 (MFD) and S3 (CCP) werexperienced urologic surgeons; the former had no experi-nce in LS or RALS, and the latter had very limitedxperience with LS and RALS (� 5 cases). Each surgeonerformed at random 5 LS and four to 5 RALS pyelo-lasties. S3 also performed 5 OS pyeloplasties. After com-leting the 5 LS and 5 RALS pyeloplasties (which we arti-icially defined as experienced), S3 went on to perform andditional 9 LS, 9 RALS, and 4 OS pyeloplasties. Fromach of these groups, 4 animals were recovered after theurgery, and the urodynamics and the histology of the op-

Abbreviations and Acronyms

Antegrade Pmax � maximal antegrade pressureLPP � leak point pressureLS � freehand laparoscopic surgeryOS � open surgeryRALS � robotic-assisted laparoscopic surgeryRetrograde Pmax � maximal retrograde pressureUPJ � ureteropelvic junction

rated UPJs were evaluated 2 to 4 weeks postoperatively. b

urgical techniquehe animals were anesthetized with telazol (4.4 to 6.6g/kg intramuscular), xylazine (1.1 to 2.2 mg/kg intra-uscular), and atropine (0.04 mg/kg intramuscular) andere maintained under anesthesia with isoflurane (1% to%) mixed with 1 to 2 L oxygen administered through anndotracheal tube. Postoperative analgesics of banamine1.1 mg/kg intramuscular or IV, immediately after surgery)nd a transdermal fentanyl patch (1 to 4 �g/kg, for2 hours) were given to those animals that underwent sur-ival surgery.

The kidney was approached retroperitoneally in the OSroup through a dorsal lumbotomy and transperitoneallyn the LS and RALS groups using the standard three-trocarechnique. We used the da Vinci Surgical System (Intuitiveurgical). With the OS group, a 5-cm incision was maderom the junction of the spine with the last rib toward thembilicus. With the laparoscopic groups, an incision wasade in the umbilicus, and access into the peritoneum was

chieved using the Veress technique. A pneumoperitoneumas then created using CO2 insufflation to a maximal pres-

ure of 12 mmHg. A 12-mm trocar was then placed throughhe Veress sheath, and the laparoscope was placed through theort. Two additional 8-mm ports were placed, one 4 cmbove and the other 2 cm to the left of the camera port,hrough which the robotic or standard laparoscopic instru-ents were placed.In all the groups, the kidney was first mobilized from the

urrounding fat, and the ureter identified. By tracing thereter toward the kidney, the renal pelvis was next identi-ied. In the laparoscopic groups, a 3-0 polydioxanone hold-ng suture (a “hitch” stitch) was advanced through the skin,hrough the renal pelvis, and back out through the skin.8

his allowed the pelvis to be suspended from the renal fossaor easier suturing. Next, the ureter was transected from theenal pelvis at the UPJ. The renal pelvis was then spatulatededially and the ureter laterally for approximately 1 cm.he ureter was reanastomosed to the renal pelvis with a

unning 5.0-polyglactin suture. The anastomotic time waseasured from the beginning to the completion of the

uturing. Of note, the time required to expose the kidney,lace the ports, and dock the robotic system were not in-luded in this measurement. No stents were used in any ofhe procedures. No specialized laparoscopic instruments orideo system was used in the LS group.

valuating the quality of the suture anastomosist the end of the procedure, while the animal was stillnder anesthesia, the anastomosis was evaluated for waterightness and patency, using retrograde and antegrade pres-ure assessments. The ureter was isolated 1 cm from the
ladder and cannulated with an 18-gauge angiocatheter

(f3mnarTsac

Ffr

csmmmmtbmk

tdp

ut vo

578 Passerotti et al Outcomes in Performing Laparoscopic Pyeloplasty J Am Coll Surg

Fig. 1). The collecting system was filled in a retrogradeashion with saline and methylene blue at a constant rate of

mL per minute using an infusion pump. All pressureeasurement was performed through the transducer con-

ected at the infusion port. Baseline pressure was measurednd then after each minute for 5 minutes. The maximaletrograde pressure (Retrograde Pmax) was determined.he volume and pressure (LPP) at which the anastomo-

is began to leak were also recorded. Measurements werelso obtained on the unoperated, contralateral kidney asontrols.

Antegrade pressure assessment was performed next.irst, the angiocatheter was removed and the ureter left toreely drain (Fig. 1). The angiocatheter was placed into the

Figure 1. Retrograde and antegrade flow asgrade flow assessment, the ureter was isolat18-gauge angiocatheter. The collecting systemmethylene blue at a constant rate of 3 mLmeasurement was performed through the trapressure was measured and then after each(LPP) at which the anastomosis began to leexhibit a high LPP, volume at leak, and maximassessment, the angiocatheter was placed ianastomosis, and the ureter left to freely draifashion with saline and methylene blue at a cpump. Baseline pressure was measured andof fluid exiting from the ureter distal to the anabeginning to the end of the antegrade infusmaximal antegrade pressure and a high outp

enal pelvis 1 cm above the apex of the anastomosis. The I

ollecting system was filled in an antegrade fashion withaline and methylene blue at a constant rate of 3 mL perinute using an infusion pump. Baseline pressure waseasured and then after each minute for 5 minutes. Theaximal antegrade pressure (antegrade Pmax) was deter-ined. The amount of fluid exiting from the ureter distal

o the anastomosis (output volume) was measured from theeginning to the end of the antegrade infusion. Measure-ents were also obtained on the unoperated, contralateral

idney as controls.An overall score for the quality of the anastomosis was

hen calculated for each procedure. Each of the parametersescribed previously except for Antegrade Pmax was ex-ressed as a fraction of the median value in the OS group.

ent of the reconstructed UPJ. In the retro-cm from the bladder and cannulated with anfilled in a retrograde fashion with saline andinute using an infusion pump. All pressure

cer connected at the infusion port. Baselinete for 5 minutes. The volume and pressureere recorded. A patent anastomosis wouldrograde pressure achieved. In antegrade flowhe renal pelvis 1 cm above the apex of the

collecting system was filled in an antegradent rate of 3 mL per minute using an infusion

after each minute for 5 minutes. The amountosis (output volume) was measured from theA patent anastomosis would exhibit a lowlume.

sessmed 1

wasper mnsduminuak wal retnto tn. Theonstathenstomion.

n contrast to all the other parameters, a high-quality

aoepftttp

HT1af

twmesoescIpp(uiwgdwbaepftctlOljtcrtt

SICrcIoCv

SAecdLapRewLvfLbiwonm

RAAtCfniStSs1StSdg


nastomosis would have a low Antegrade Pmax (ie, nobstruction); consequently, the fraction for this param-ter was inverted to make it concordant with the otherarameters. When the Antegrade Pmax � 0, then theraction was defined as 1. The sum of the fractions washen divided by 5. The closer the overall score was to 1,he more similar the quality of the anastomosis was tohat of OS. We arbitrarily used this score to compare allarameters simultaneously.

istologic evaluationhe histology of the operated UPJ was evaluated in the last2 animals (4 in each group all performed by S3 15 daysfter surgery). The specimens were first fixed in 10%ormalin.

Five-�m sections were then obtained through the anas-omosis (marked by the sutures). The sections were stainedith hematoxylin and eosin, and the length of the anasto-osis and the diameter of the UPJ were measured. The

xtent of the inflammatory reaction around the anastomo-is was also evaluated by measuring the maximal thicknessf the UPJ wall. The amount of collagen deposition wasvaluated using a picrosirius-hematoxylin method. In brief,ections were deparaffinized, hydrated to water, and in-ubated for 20 minutes at 37° C in 0.5% papain (PapainV F VIII; Difco Laboratories) dissolved in 0.02 Mhosphate buffer, containing 5 mmol/L sodium bisul-hite and 0.5 mmol/L ethylenediaminetetraacetic acidEDTA), pH 4.7. Sections were then stained for 30 min-tes in a 0.2% Sirius red solution (Direct Red 80; Aldrich)

n saturated picric acid, followed by a rapid rinse in tapater, dehydrated, and mounted in synthetic resin. Colla-en, which is rich in basic amino acids, reacts with acidicyes. Sirius red is an elongated dye molecule that reactsith collagen and promotes an enhancement of its normalirefringence because of the fact that many dye moleculesre aligned parallel with the long axis of each collagen mol-cule. The enhancement of birefringence promoted by theicrosirius staining under polarized microscopy is specificor collagenous structures composed of aggregates of orien-ated molecules. Papain digestion hydrolyzes the proteinore of proteoglycans, extracting these compounds fromissues without affecting its collagen content. To study col-agen organization, polarization was obtained using an

lympus BX 51 microscope with two Polaroid filters, 1ocated below the condenser and the other above the ob-ective lens. Mature collagen (type I) strongly reacts withhe Sirius red dye, resulting in bright red coloration. Inontrast, collagen type III weakly binds with the Siriused dye and, as a result, appears green-yellow. For quan-itative analysis, photographic images were obtained with
he Olympus Q video camera attached to the microscope. p
tandard correction for shading was performed. Using themage Pro Plus – Optimas 6.0 for Windows (Mediaybernetics), collagen composition was determined in 10

andom fields per sample by measuring the amount of redolor fibers (type I collagen) and green-yellow fibers (typeII collagen). This measurement of collagen density wasbtained from the lamina propria to the muscularus layer.ollagen associated with the basal membrane and blood

essels was excluded in the density analysis.

tatistical analysis2-way analysis of variance (ANOVA) test was used to

valuate for statistical significance (p � 0.05) whenomparing the anastomotic time of surgeons indepen-ent of technique and when comparing the RALS andS groups independent of surgeon. A linear regressionnalysis was used to determine if there was serial im-rovement in operative time by each surgeon using theALS and LS techniques; a p value � 0.05 was consid-red statistically significant. A Kruskall-Wallis analysisas performed to determine statistical significance ofPP, volume at leak, retrograde pressure, and outputolume, using the appropriate p value given by the Bon-erroni’s correction for each parameter analyzed. ForPP, p � 0.007 was considered statistically significantased on the Bonferroni’s correction for this parametern performing 11 comparisons; the appropriate p valueas used for the other measurements. A generalizationf Fisher’s exact test was used to evaluate statistical sig-ificance of antegrade pressure measurement, becauseost of the results within this parameter were zero.

ESULTSnastomotic time and the learning curvenastomotic time for the OS group was similar between

he surgeons during the first few cases (data not shown).onsequently, we combined those data and had S3 per-

orm all the pyeloplasties in the OS group to minimize theumber of animals used for the study. Figures 2A to 2D

llustrate the anastomotic time for S1, S2, the first 5 cases of3, and for all of S3’s cases, respectively. When analyzinghe operative time independent of the techniques, S1 and2 had similar operative times (p � 0.50); S3 had thehortest operative time (16.3 � 5 minutes less than S1 and5.0 � 5.1 minute less than S2; p � 0.01). After 5 cases,3’s mean operative time was 15.4 � 5.0 minutes less thanhe times for his first 5 cases, 31.7 � 5.0 minutes less than1’s time, and 30.4 � 5.1 minutes less than S2’s, indepen-ent of technique (all with p � 0.01). These findings sug-ested that independent of technique, S1 and S2 had com-
arable anastomotic times despite the difference in their

mfqL

st0Cmcar2t3far

of

Tt

S

SSS

AS

*u†

h cas


edical training. As expected, more experience (ie, per-orming more procedures) greatly decreased the time re-uired to perform the anastomosis with both the RALS andS techniques.

able 1. Analysis of Anastomotic Time Using Open, Conven-ional Laparoscopic, and Robotic-Assisted Techniques

urgeon

Mean anastomotic time, min � SD

Open LaparoscopicRobotic-assisted

laparoscopy

1 — 73.0 � 11.0 44.0 � 17.2 (p � 0.01)*†

2 — 65.6 � 19.2 48.3 � 11.9 (p � 0.16)*3 (first 5

cases) 22.3 � 4.6 54.0 � 8.2 31.8 � 3.4 (p � 0.01)*†

ll surgeons 64.2 � 14.9 40.9 � 13.4 (p � 0.01)*†

3 (last 5cases) — 31.4 � 3.4 22.4 � 4.1 (p � 0.52)*

p values comparing laparoscopic versus robotic-assisted laparoscopy groupsing a two-way ANOVA test.

Figure 2. Ureteropelvic anastomosis time of the first 5 cases using rowithout any robotic-assisted laparoscopic surgery (RALS) or freehand lawithout any RALS or LS experience; (C) S3, an experienced urologic suall of his cases. The p values are calculated based on comparing eac

cp � 0.05.

The mean LS and RALS anastomotic times for eachurgeon are listed inTable 1; the average RALS anastomoticime was statistically less than that of LS for S1 and S3 (p �.01) and trended toward statistical significance for S2.ombining all the surgeons’ times for each group, theean suture time for RALS was 40.9 � 13.4 minutes,

ompared with 64.2 � 14.9 minutes with LS (p � 0.01)nd 22.4 � 4.1 minutes with OS (p � 0.01). With expe-ience (S3’s last 5 cases), the mean suture time for RALS of2.4 � 4.1 minutes approached that of OS (p � 0.80);he mean operative time for LS also improved to1.4 � 3.4 minutes (p � 0.10 compared with OS). Theseindings suggested that independent of baseline experience,ll surgeons were able to perform the anastomosis moreapidly using RALS than with LS.

To evaluate individual improvement case to case for eachf the techniques, a linear regression analysis was per-ormed. With RALS, S1 and S3 had a statistically signifi-

assistance or freehand laparoscopy by (A) S1, a nonurologic surgeoncopic surgery (LS) experience; (B) S2, an experienced urologic surgeonwith minimal RALS and LS experience (first 5 cases); and (D) S3 for

e to the next using linear regression analysis. OS, open surgery.

boticparosrgeon

ant decrease in the anastomotic time from case to case,

witwttcapw

QiRlkPsLagtoipwpbpatriaRslgR

(dLrg

akwitgLbdwcR2foweg(

vaOgwLigqt�s

TF

F

L

V

R

A

O

M

*A ; RALS


ith the less experienced surgeon (S1) having the greatestmprovement (Fig. 2). S2 also had a decrease in the anas-omotic time from case to case, but this only trended to-ard statistical significance. In comparison, there was a

rend toward serial improvement in the LS anastomoticime by S2 and S3; S1 had no statistically significanthange. These findings suggested that the learning curvessociated with RALS appeared to be more favorable com-ared with that of LS, with decreasing anastomotic timeithin fewer numbers of cases.

uality of the anastomosismmediately postsurgeryetrograde flow measurement was performed to assess

eakage at the anastomosis (Fig. 1). In the unoperatedidney, the maximal retrograde pressure (Retrogrademax) achieved was 28.1 � 10.9 cm H20, with no mea-urable LPP and volume at leak. For all surgeons, thePP, volume at leak, and maximal Retrograde Pmaxchieved were higher in the RALS compared with the LSroup (p � 0.01, Table 2) but were statistically lowerhan those for the unoperated kidney (p � 0.01). All 3f these parameters in the RALS group were not signif-cantly different between S1, S2, and S3 (first 5 cases;� 0.20) but did become significant when comparedith S3’s last 5 cases (p � 0.01). All three of thesearameters in the LS group were significantly differentetween S1, S2, S3 (first 5 cases) and S3 (last 5 cases;� 0.01). For S3‘s last 5 cases, the LPP, volume at leak,

nd Retrograde Pmax in the RALS group were similar tohose of the OS group (p � 0.30); those in the LS groupemained significantly different (p � 0.05). These find-ngs suggested that immediately after operation, all thenastomoses leaked. But the anastomoses done withALS compared with LS had a higher leak point pres-

ure and achieved a greater volume and pressure beforeeaking, even when performed by inexperienced sur-eons. With experience, all of these parameters in the

able 2. Retrograde and Antegrade Flow Measurements inreehand Laparoscopy by Each of the 3 Surgeons (S1, S2,

low measurements Unoperated

S1

LS RALS LS

PP, cm H20 NA 3.4 � 3.1 6.8 � 4.5 0.1 � 0.2

olume at leak, mL NA 3.6 � 1.3 7.2 � 2.7 2.2 � 1.1

etrograde Pmax, cm H20 28.1 � 10.9 3.7 � 2.9 8.0 � 5.9 0.5 � 1.1

ntegrade Pmax, cm H20 0 0.4 � 0.8 0 3.0 � 4.3

utput volume, % 100 47 � 39 89 � 24 20 � 45

ean quality score, % NA 39 � 15 61 � 12* 25 � 10

p value � 0.01 comparing LS versus RALS using 2-sided ANOVA test.ntegrade Pmax, maximal antegrade pressure; LPP, leak point pressure; LS, laparoscopy

ALS group approached those of the OS group R

p � 0.30). In contrast, more practice (more proce-ures) was required to improve these parameters in theS group; even after 10 cases, all 3 of these parametersemained statistically different from those of the OSroup (p � 0.01).

In addition, antegrade flow measurement was performed tossess patency of the anastomosis (Fig. 1). In the unoperatedidney, the maximal antegrade pressure (Antegrade Pmax)as 0, and the output volume was 100% of what was

nfused, because the distal ureter was left to drain freely andhere was unrestricted flow through the UPJ. For all sur-eons, the Antegrade Pmax was low for both the RALS andS groups (p � 0.20 when compared with each other, Ta-le 2). This indicates that there was no obstruction createduring the pyeloplasty (Table 2). But the output volumeas significantly higher (p � 0.04) in the RALS group

ompared with the LS group, indicating less leakage in theALS. With S2, elevated Antegrade Pmax was observed inanimals in the LS group (second and fourth) and in the

irst animal in the RALS group, suggesting a degree ofbstruction. With S1, a slight increase in Antegrade Pmaxas observed in the third case in the RALS. With furtherxperience, all parameters in both the RALS and LSroups improved and approached those of the OS groupp � 0.40, Table 2).

To simultaneously take into account all the urodynamicariables, an overall quality score was calculated. This over-ll quality score was expressed as a ratio to the score in theS group (the gold standard in pyeloplasty). For each sur-

eon, the overall score for the quality of the anastomosisas higher in the RALS group compared with that in theS group (Fig. 3, A to C). But the score improved with

ncreasing number of cases in both the RALS and LSroups and approached that of OS (Fig. 3D). The meanuality score for the RALS group was statistically higherhan that of the LS group for S1, S2, and S3 (first 5 cases; p

0.01, Table 2). For S3 (last 5 cases), the mean qualitycore was not statistically different between the LS and

noperated and Operated Kidney Using Robotic-Assisted orS3)

S3

First 5 cases Last 5 cases

RALS LS RALS Open LS RALS

4 � 11.7 2.9 � 1.7 13.1 � 3.5 19.1 � 3.4 11.9 � 0.7 17.6 � 6.5

8 � 2.9 4.2 � 1.6 12 � 0 10.2 � 1.6 12 � 0 9.6 � 2.5

6 � 12.2 9.0 � 5.6 14.1 � 3.3 24.4 � 5.4 12.9 � 0.5 25.8 � 6.4

3 � 2.7 0.2 � 0.3 0.1 � 0.1 0.2 � 0.5 0 0

100 73 � 43 100 100 100 100

4 � 18* 54 � 16 84 � 5* NA 87 � 3 92 � 8

, robotic-assisted laparoscopy; Retrograde Pmax, maximal retrograde pressure.

the Uand

S2

11.

9.

13.

1.

7

ALS groups.

QWmohatesiT

vamlgcbdg

tfeeLoglsLcacl

TtLM

LVRAOL

Alr

tomo


uality of the anastomosis 15 days postsurgerye performed retrograde and antegrade flow measure-ents and histologic evaluation of the UPJ 15 days post-

peratively to assess the quality of the anastomosis withealing. To control for variability between surgeons, wessessed only those animals operated on by S3 and only inhe last 4 animals in each group to control for the effects ofxperience. We observed that no animals had any demon-trable leak. Consequently, the Retrograde Pmax was sim-lar between the RALS, LS, and OS groups (p � 0.10,able 3). With the Antegrade Pmax of 0 and the output

able 3. Retrograde and Antegrade Flow Measurements ofhe Ureteropelvic Anastomosis 15 Days after Freehandaparoscopic, Robotic-Assisted, and Open Surgery by S3easurements LS RALS Open

PP, cm H20 0 0 0olume at leak, mL 0 0 0etrograde Pmax, cm H20 30.4 � 4.6 29.7 � 3.9 22.5 � 4.3ntegrade Pmax, cm H20 0 0 0utput volume, % 100 100 100ength of theanastomosis, cm 1.6 � 0.5 1.4 � 0.3 1.7 � 0.4

ntegrade Pmax, maximal antegrade pressure; LPP, leak point pressure; LS,aparoscopy; RALS, robotic-assisted laparoscopy; Retrograde Pmax, maximal

Figure 3. Overall quality score of ureteropelvic anastomosis of theby each of the 3 surgeons: (A) S1, (B) S2, (C) S3, and (D) the last 5was expressed as a fraction of the median value in the open surgeThe closer the overall score is to one, then the quality of the anas

aetrograde pressure.

olume being 100% for all 3 groups, the patency of thenastomosis was confirmed on the antegrade flow measure-ent (Table 3). On microscopic examination, the mean

engths of the anastomoses were 1.67 � 0.40 cm in the OSroup, 1.62 � 0.50 cm in the LS group, and 1.47 � 0.30m in the RALS group; there was no statistical differenceetween the groups (p � 0.80). Similarly, there was noifference in the diameter of the UPJ between the threeroups.

Hematoxylin and eosin staining demonstrated typicalhickening of the ureteral wall at the anastomosis, primarilyrom collagen deposition with inflammatory infiltrates anddema (Fig. 4). This inflammatory reaction was less appar-nt in the RALS group (6.3 � 1.7 mm) compared with theS group (8.8 � 3.0 mm; p � 0.01) but was similar to thatf the OS group (5.1 � 2.3 mm; p � 0.50). The total colla-en content at the anastomotic site trended toward being theeast in the RALS group, but this did not achieve statisticalignificance (RALS � 3.1 � 0.5 �m, OS � 4.4 � 0.5 �m,S � 4.8 � 1.5 �m; p � 0.14; Fig. 5). With regard to theollagen types, there was no statistical difference in themount of collagen type I in the open group (0.8 � 0.2 �m)ompared with the RALS group (1.0 � 0.5 �m) and theaparoscopic group (1.2 � 0.8 �m; p � 0.15). But the

cases using robotic-assisted (RALS) or freehand laparoscopy (LS)s by S3. Each of the retrograde and antegrade flow measurements) group. Overall score is the sum of all parameters divided by five.

sis is similar to that of OS.

first 5case

ry (OS

mount of collagen type III was significantly less in the RALS

g�

DTahoacctlLtttapiac

ttiomamaac

qgtpopitlsgegw

sttcstsri

pi1tepf

r in a


roup (2.3 � 0.5 �m) than in the open group (3.2 � 0.5m) and the laparoscopic group (3.5 � 2.0 �m; p � 0.01).

ISCUSSIONhis is the first study to directly evaluate the learning curvend the quality of the anastomosis associated with free-and and robotic-assisted laparoscopic suturing. It dem-nstrated that suture anastomosis of the UPJ with roboticssistance can be accomplished in a shorter amount of timeompared with freehand laparoscopy and required fewerases to approach the times of open surgery. The quality ofhe anastomosis (as measured by patency and minimaleakage) was better in the RALS group compared with theS group. But it should be noted that with practice, bothhe anastomotic time and the quality of the anastomosis inhe LS group also approached those of open surgery. His-ologic findings indicated that the patency was similar withll three techniques, but the amount of collagen III ap-eared to be less in the RALS group. Together, these find-ngs suggest that for the inexperienced surgeon, roboticssistance had distinct advantages over freehand laparos-opy when performing a pyeloplasty.

Laparoscopic suturing is a technically demanding skillhat traditionally has required a significant amount of timeo master. In some cases, proficiency may never be achievedn a reasonable amount of time, because of the infrequencyf patients. The need for improving suturing ability ininimally invasive surgery has required development of

nimal models to test these new techniques.10,11 Surgicalodels performed in pigs have been well established and

re invaluable in laparoscopic training.12,13 Despite thevailability of these animal models, any modification in

Figure 4. Hematoxylin and eosin staining of thsection. Thickening in the ureteral wall at theincreased collagen deposition without significance of the ureteropelvic junction was simila

onventional laparoscopic technique that reduced the re- r

uired learning time to achieve proficiency would be ofreat benefit. In the surgical management of UPJ obstruc-ion, it has been shown previously that laparoscopic pyelo-lasty in children is associated with reduced morbidity andutcomes comparable to those in open surgery.14,15 But toerform this procedure with the same efficiency and qual-ty as that of open pyeloplasty requires advanced and ex-ensive laparoscopic suturing experience. Consequently,aparoscopic pyeloplasty in children has not gained wide-pread popularity at many institutions. Our findings sug-est that for the inexperienced surgeon, laparoscopic py-loplasty can be performed in a timely manner and be ofood quality using robotic-assisted techniques, achievableithin a small number of cases (ie, short learning curve).We chose pyeloplasty as our surgical model in assessing

uturing skills because of its inherent advantages, includinghe relatively constant anatomy and renal pelvis and ure-eral caliber, which minimize anatomic variations that mayonfound the results. To confirm this premise, we mea-ured the length of the anastomosis performed in each ofhe groups and found that it was similar in all three. Con-equently, the anastomotic time could be compared di-ectly, because the amount of suturing required was similarn all three groups.

In our study we measured only the time required toerform the suturing. The overall procedure time, whichncluded that required for setup, was an average of0 minutes longer in the RALS group compared withhat for the LS and OS groups. Obviously, if we definefficiency as the time required to perform the entirerocedure, RALS is less favorable than open surgery andreehand laparoscopy. But in this study we chose to cor-

teropelvic anastomosis (10�) in longitudinaltomosis was observed, primarily because offlammatory infiltrates or edema. This appear-ll three groups.

e ureanas

ant in

elate efficiency with the level of difficulty in performing

tadsdhucpjtttisoagtngeInbil

aIaamomisamceaRgtcw

tws

FmsIgmigroups. The suture can be seen as black and white fibers (S).


he pyeloplasty. Hence, we elected to evaluate only thenastomotic time, a more objective measurement of theifficulty associated with laparoscopic suturing. Theurgeons in our study indicated that it was much moreifficult for them to perform the pyeloplasty with free-and laparoscopy. They expressed frustration in manip-lating the needle and in knot-tying and had a signifi-ant degree of hand, arm, and shoulder discomfort aftererforming the procedure. But these findings were sub-

ective. The anastomotic time represented a more objec-ive measurement of the degree of difficulty, allowing uso more directly compare the three techniques. In addi-ion, it allowed us to monitor improvement in perform-ng the pyeloplasty with these various techniques as theurgeons acquired more experience. In our study webserved that the anastomotic time of the RALS group,t the onset, was significantly lower than that of the LSroup. Consequently, the learning curve associated withhe RALS group approaches that of OS within a smallumber of cases compared with the LS group. This sug-ests that learning to suture laparoscopically is inher-ntly less difficult with the use of robotic assistance.nterestingly, a lack of urologic surgical experience didot have an impact on the ability of the surgeon to learnoth robotic-assisted and freehand laparoscopic sutur-ng. The inexperienced urologic surgeon had a similarearning curve to that of the experienced surgeon.

Being able to suture in a timely manner should not comet the sacrifice of performing a high-quality anastomosis.n this study, we assessed patency and leakage immediatelynd 15 days after surgery. Our rationale for evaluating thenastomosis early on was that if there was a leak, it wouldost likely seal after several days. So, immediate evaluation

f the UPJ would be informative. Our urodynamic assess-ent of the UPJ was performed at a flow rate approximat-

ng that of normal urine output16,17to replicate physiologictress on the anastomosis. Based on the retrograde andntegrade flow measurements, we found that the anasto-oses in the RALS group were patent and had less leakage

ompared with those in the LS group and, with increasedxperience, were similar to those in the OS group. As withnastomotic time, the quality of the anastomosis in theALS group from the onset was higher than that of the LSroup. But with enough practice, the anastomotic time andhe quality of the anastomoses in the LS group also becameomparable to those of the OS group; this is consistentith clinical experience.Urodynamic and histologic evaluation done 15 days af-

er surgery similarly demonstrated that all the anastomosesere patent. Interestingly, the degree of fibrosis, as mea-

igure 5. Picrosirius staining of the ureteropelvic junction anasto-osis (60�) in the (A) open, (B) robotic-assisted laparoscopic

urgery (RALS), and (C) freehand laparoscopic groups. Collagen typeis designated red and type III as green. Note in the laparoscopicroup there is an increased density of collagen around the anasto-osis compared with the other two groups. Collagen type III (green)

s the highest in this group compared with the open and the RALS

ured by the amount of collagen deposition, was less in the

RlTaipsfmq

eolaidtawtadpnRudpfnsrar

saeolpufec

ASA

A

D

C

AKiS

R

1

1

1

1

1

1

1


ALS group compared with the LS and OS groups. Col-agen is an essential component of the extracellular matrix.ypes I and III are the interstitial or fibrillar collagens andre most abundant. Several studies have demonstrated thatncreased collagen type III is associated with loss of com-liance18 and contractility19-21 in the urinary system. Wepeculated that the increased optical magnification allowedor more precise tissue handling with RALS. This, in turn,ay have resulted in less trauma to the tissue and conse-

uently less fibrotic reaction.The findings in this study suggest that for the inexperi-

nced surgeon, laparoscopic suturing is easier to learn at thenset with robotic assistance compared with freehandaparoscopy, without compromising the quality of thenastomosis. But we also recognize that performing sutur-ng is only one component of the operation. Success is alsoependent on other factors, such as accurate placement ofhe ports, which were not assessed during this study. Inddition, the ease of laparoscopic surgery may be improvedith the use of specialized instruments that could improve

he efficiency of performing certain laparoscopic skills suchs tying a knot. In addition, we recognize that RALS hasistinct disadvantages. The longer procedure time, the ex-ense of the robotic system, and the need for highly trainedurses and assistants can impede the more general use ofALS despite its potential benefits. Nevertheless, we spec-late that with increased use of robotic assistance, theseisadvantages can be minimized. With practice and ex-erience, more nurses and surgeons can become moreamiliar with the robotic system and procedure, elimi-ating the need for special personnel and decreasing theetup and overall procedure time. In time, this couldesult in decreasing expenses (eg, operating room timend hospitalization) that may offset the initial cost of theobotic system.

In conclusion, our findings, the first in an objectivelytudied animal model, suggest that RALS pyeloplasty isssociated with a short learning curve, with the mean op-rative time approaching that of OS within a short numberf cases. In contrast, LS pyeloplasty is associated with aonger learning curve than OS. The water tightness andatency of the suture anastomosis can be achieved rapidlysing RALS and are comparable to those of OS. Theseindings suggest that over time RALS may be performed,ven by inexperienced surgeons, with efficiency and resultsomparable to open surgical approaches.

uthor Contributionstudy conception and design: CC Passerotti, Nguyencquisition of data: CC Passerotti, AMS Passerotti,
Dall’Oglio 1
nalysis and interpretation of data: CC Passerotti, AMSPasserotti, Leite, Nunes, Srougi, Nguyenrafting of manuscript: CC Passerotti, AMS Passerotti,Retik, Nguyen

ritical revision: CC Passerotti, AMS Passerotti, Srougi,Retik, Nguyen

cknowledgment: We thank Dr Arthur Nedder, Arthurelly, Alicia Frank, Kris Guffa, and Selby Lebert for their help

n caring for the animals. In addition, we thank Dr Adrianaanudo for the extensive statistical analysis.

EFERENCES

1. Buckingham RA, Buckingham RO. Robots in operating the-atres. BMJ 1995;311:1479–1482.

2. Peters CA. Robotically assisted surgery in pediatric urology. UrolClin North Am 2004;31:743–752.

3. De Ugarte DA, Etzioni DA, Gracia C, et al. Robotic surgery andresident training. Surg Endosc 2003;17:960–963.

4. Hemal AK, Menon M. Robotics in urology. Curr Opin Urol2004;14:89–93.

5. Bonnard A, Fouquet V, Carricaburu E, et al. Retroperitoneallaparoscopic versus open pyeloplasty in children. J Urol 2005;173:1710–1713; discussion 1713.

6. Reed MJ, Williams MP. Open pyeloplasty in children: experi-ence with an improved stenting technique. Urol Int 2003;71:201–203; discussion 203.

7. Atug F, Woods M, Burgess SV, et al. Robotic assisted laparo-scopic pyeloplasty in children. J Urol 2005;174:1440–1442.

8. Lee RS, Retik AB, Borer JG, et al. Pediatric robot assistedlaparoscopic dismembered pyeloplasty: comparison with acohort of open surgery. J Urol 2006;175:683–687; discussion687.

9. Olsen LH, Rawashdeh YF, Jorgensen TM. Pediatric robot as-sisted retroperitoneoscopic pyeloplasty: a 5-year experience.J Urol 2007;178:2137–2141; discussion 2141.

0. Chiu AW, Lin CH, Huan SK, et al. Creation of ureteropelvicjunction obstruction and its correction by chemical glue-assistedlaparoscopic dismembered pyeloplasty. J Endourol 2003;17:23–28.

1. Shumalinsky D, Lobik L, Cytron S, et al. Laparoscopic lasersoldering for repair of ureteropelvic junction obstruction in theporcine model. J Endourol 2004;18:177–181.

2. Clayman RV. Surgical robotics: impact of motion scaling on taskperformance. J Urol 2005;174:953.

3. Hubert J, Feuillu B, Mangin P, et al. Laparoscopic computer-assisted pyeloplasty: the results of experimental surgery in pigs.BJU Int 2003;92:437–440.

4. Reddy M, Nerli RB, Bashetty R, et al. Laparoscopic dismem-bered pyeloplasty in children. J Urol 2005;174:700–702.

5. Tan HL. Laparoscopic Anderson-Hynes dismembered pyelo-plasty in children. J Urol 1999;162:1045–1047; discussion1048.

6. Frokiaer J, Tofft HP, Mortensen J, et al. Peristaltic activity in thenormal renal pelvis of the pig during standardized perfusions.Urol Res 1985;13:249–252.

7. Mortensen J, Djurhuus JC, Laursen H, et al. The relationship

1

1

2

2


between pressure and flow in the normal pig renal pelvis. Anexperimental study of the range of normal pressures. ScandJ Urol Nephrol 1983;17:369–372.

8. Baskin LS, Constantinescu S, Duckett JW, et al. Type III colla-gen decreases in normal fetal bovine bladder development.J Urol 1994;152:688–691.

9. Murakumo M, Nonomura K, Yamashita T, et al. Structural
changes of collagen components and diminution of nerves in
congenital ureteropelvic junction obstruction. J Urol1997;157:1963–1968.

0. Sergi F, Flammia GP, Alcini A, et al. Collagen changes in theureteropelvic junction after failed antegrade endopyelotomy. JEndourol 2007;21:103–107.

1. Yoon JY, Kim JC, Hwang TK, et al. Collagen studies for pedi-atric ureteropelvic junction obstruction. Urology 1998;52:494–
497; discussion 497–498.
NOTICE TO AUTHORSSurgeon at Work articles submitted after October 31, 2007will no longer be published in the print Journal. They will be

published ONLINE ONLY and referenced in the table ofcontents with an �e-page� number. We encourage

multimedia submissions (see Instructions for Authors atwww.journalacs.org)

Comparing the Quality of the Suture Anastomosis and the Learning Curves Associated with Performing...

Documents

Transcript of Comparing the Quality of the Suture Anastomosis and the Learning Curves Associated with Performing...