Endoscopy 2012; 44(06): 596-604
DOI: 10.1055/s-0031-1291790
Original article
© Georg Thieme Verlag KG Stuttgart · New York

Standard diagnostic laparoscopy is superior to NOTES approaches: results of a blinded, randomized controlled porcine study

D. von Renteln
1   Department of Interdisciplinary Endoscopy, Hamburg-Eppendorf University Hospital, Hamburg, Germany
,
T. E. Gutmann
2   Mannheim University Hospital, Heidelberg University Medical School, Mannheim, Germany
,
A. Schmidt
3   Department of Gastroenterology, Hepatology and Oncology, Ludwigsburg Hospital, Ludwigsburg, Germany
,
M. C. Vassiliou
4   Department of Surgery, McGill University Health Centre, Montreal, Quebec, Canada
,
H. U. Rudolph
5   Department of Surgery, Mannheim University Hospital, Mannheim, Germany
,
K. Caca
3   Department of Gastroenterology, Hepatology and Oncology, Ludwigsburg Hospital, Ludwigsburg, Germany
› Author Affiliations
Further Information

Corresponding author

D. von Renteln, MD
Department of Interdisciplinary Endoscopy
University Hospital Hamburg-Eppendorf
Martinistr. 52
20246 Hamburg
Germany   
Fax: +49-040-741040004   

Publication History

submitted 21 July 2011

accepted after revision 09 January 2012

Publication Date:
08 March 2012 (online)

 

Background and study aim: The use of transluminal endoscopic access via the stomach or colon for flexible diagnostic peritoneoscopy has been proposed, although the diagnostic value of the technique has not yet been fully clarified. In this animal trial, the two main natural orifice transluminal endoscopic surgery (NOTES) approaches – transgastric (TG) and transcolonic (TC) – were compared with standard transabdominal access using both rigid (TAR) and flexible instruments (TAF) for diagnostic laparoscopy.

Methods: A total of 48 peritoneoscopies were performed using two randomly assigned approaches in 24 anesthetized pigs. The ability of the examinations to detect 576 electrocautery markings simulating intraperitoneal metastases, to achieve complete organ visualization, and to simulate organ biopsies was analyzed.

Results: Sensitivities for the detection of lesions were 78.5 %, 59.7 %, 48.6 %, and 38.9 % for TAR, TAF, TC, and TG, respectively; standard laparoscopy was superior to all other approaches (P < 0.01). Among the NOTES approaches, TC was superior for examining the upper abdomen (P = 0.03). Complete organ visualization was better with the transabdominal approach (visual analogue scale TAR 7.15, TAF 6.71) than with the NOTES access routes (TC 5.07, TG 4.35); standard rigid laparoscopy was superior to both NOTES approaches (P < 0.01). Organ biopsy simulation was possible in 87 %, 85 %, 72 %, and 65 % of cases with TAR, TAF, TC, and TG, respectively. Standard rigid laparoscopy was again superior to both NOTES approaches (TAR vs. TC, P = 0.03; TAR vs. TG, P < 0.01).

Conclusions: In this experimental trial, rigid standard laparoscopy provided better organ visualization, better lesion detection, and better biopsy capability than the transgastric and transcolonic NOTES approaches. In its current form, NOTES appears to be unsuitable for diagnostic laparoscopy.


Introduction

Diagnostic laparoscopy allows minimally invasive staging of intra-abdominal cancers, such as esophageal, gastric, biliary, and pancreatic cancer [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11]. Less frequent indications include hepatocellular carcinoma, colorectal cancer, and lymphoma staging [12] [13] [14] [15] [16] [17] [18]. Diagnostic laparoscopy for cancer staging is commonly indicated in candidates for curative surgical resection [1] [3] [5] [19] and should be used after high quality imaging has shown no evidence of distant metastases [1] [4] [5] [9]. Diagnostic laparoscopy increases the rate of detection of intraperitoneal metastases and is a better method of selecting those patients who should receive curative therapy rather than neoadjuvant or palliative treatment [5] [7] [20] [21]. In addition, its use can avoid nontherapeutic laparotomy and the associated complications [5] [6] [10] [22].

Laparoscopy forms the basis of the minimally invasive surgical approach, which has recently been further developed toward the goal of completely incisionless surgery [23] [24]. Natural orifice transluminal endoscopic surgery (NOTES) is a novel method for accessing the peritoneal cavity through the mouth, vagina, or rectum, thereby avoiding any skin incisions [23]. However, the best clinical targets for NOTES remain unclear at present, with increasing numbers of therapeutic animal trials becoming available for a variety of abdominal resection procedures [25] [26].

The initial procedure for which NOTES was tested in animals was diagnostic peritoneoscopy or laparoscopy [24] [27] [28] [29]. More recently, animal research has indicated that transgastric peritoneoscopy for the detection of intraperitoneal lesions may have limitations [30]. Nevertheless, the first applications of the method in humans have been presented, without clear evidence of the value of transluminal laparoscopy in comparison with the standard approach using rigid laparoscopy [27] [28].

A systematic, prospective, randomized study was therefore conducted to compare the two main access routes for flexible transluminal laparoscopy – transgastric NOTES (TG) and transcolonic NOTES (TC) – with the standard rigid laparoscopic transabdominal approach (TAR). To further analyze the value of the flexible endoscope used in NOTES in comparison with the mostly rigid scopes used for transabdominal access, a fourth study group was included in which flexible endoscopes were used through the standard transabdominal access route (TAF). The three main functions of diagnostic laparoscopy were analyzed in each of the four groups – namely, complete organ visualization, detection of peritoneal lesions, and the ability to perform (simulated) biopsies.


Materials and methods

Animals and pre-interventional preparation

A total of 48 acute animal experiments were performed in 24 female domestic pigs (mean weight 34 kg, range 21 – 54 kg, standard deviation [SD] 10.3) after approval had been obtained from the Animal Care and Use Committee. The animals were fasted from solid food for 48 hours before surgery, but were allowed full access to water and milk. Sedation pre-anesthesia consisted of ketamine 2 mg/kg and xylazine 2 mg/kg. General anesthesia was achieved using isoflurane, N2O, and O2, following endotracheal intubation. Pulse oximetry and electrocardiography were monitored continuously throughout the procedures, which were performed with the animals in the supine position. When the animal was under general anesthesia, gastric and colonic lavages were carried out with 1000 mL tap water using an endoscopic water-jet pump (AFU-100; Olympus, Hamburg, Germany). A Veress needle was introduced, and a 15-mmHg CO2 pneumoperitoneum was established. Four lateral 5.5-mm trocars and one paraumbilical 5.5-mm trocar (TroQ; Olympus) were introduced ([Fig. 1]). A 15-mmHg CO2 pneumoperitoneum was maintained during all procedures.

Zoom
Fig. 1 Port location for creation of lesions within the peritoneal cavity and examination by two of the four approaches. a The locations at which the one paraumbilical trocar and four lateral trocars were introduced into the anesthetized pigs. b Photograph of two of the lateral 5.5-mm trocars that were used to create the electrocautery markings positioned ready for use. c Photograph of transabdominal diagnostic peritoneoscopy with a flexible endoscope being performed via the single paraumbilical trocar.

Outcome variables

Three outcome variables were analyzed for this study.

(a) Visualization of organs

This consisted of a predefined, thorough, and systematic inspection of the right lobe, left lobe, upper and lower surfaces of the liver; gallbladder; right and left hemidiaphragms; anterior wall and greater curvature of the stomach; spleen; greater omentum; small bowel; sigmoid colon; rectum; pouch of Douglas; right and left groins; right and left ovaries; right and left uterine horns; and the right upper quadrant, left upper quadrant, right lower quadrant, and left lower quadrant of the abdominal wall. For each of these organs and areas, the observer rated the completeness of organ surface visualization on a visual analogue scale (VAS; scale 0 to 10), with 0 meaning that the organ or area could not be visualized at all (0 %) and 10 representing complete (100 %) visualization of the organ surface. To facilitate optimal exposure of all intraperitoneal organs and areas in this examination, the animal was systematically placed in the Trendelenburg, anti-Trendelenburg, 30° left, and 30° right positions. Additional positioning was facilitated as requested by the observer.


(b) Biopsy targeting simulation

To simulate the performance of a biopsy and to evaluate whether the bending of flexible instruments made it possible to access distant areas better than the rigid instrument, the investigator also attempted to touch all of the areas listed above with the instrument tip. The time allowed for these tests (a) and (b) was limited to 30 minutes.


(c) Detection of premarked lesions

To simulate the appearance of intraperitoneal metastases, 1 – 2-mm electrocautery markings were made in the peritoneal cavity using a flexible endoscope (GIF N180; Olympus) and the tip of a standard diathermic snare ( [Fig.2]). In order to cover the entire peritoneal cavity, markings were made via the four lateral laparoscopic ports in each abdominal quadrant ([Fig. 1]). These were exclusively used to create the electrocautery markings and were not used for subsequent diagnostic procedures. During marker placement, animals were positioned in the Trendelenburg, anti-Trendelenburg, and 30° left and right lateral positions in order to achieve optimized access to all intraperitoneal organs and areas.

Zoom
Fig. 2  Example lesions in the model of peritoneal metastases. a Laparoscopic view of a single electrocautery marking on the gallbladder. b Endoscopic view of a double electrocautery marking on the gallbladder. c Laparoscopic view of a single electrocautery marking on the liver. d Endoscopic view of a single electrocautery marking on the liver. e Laparoscopic view of a single electrocautery marking on the abdominal wall. f Endoscopic view of a double electrocautery marking on the abdominal wall.

As each animal underwent examination twice, using a laparoscopic and a NOTES technique, the second round required the detection of double markings that were made separately ([Fig. 3]). The electrocautery markings were evenly distributed within the quadrants of the peritoneal cavity at each round, leaving three detectable electrocautery markings in each quadrant. Placement was based on likely targets for metastatic disease in upper gastrointestinal, hepatic, and biliopancreatic malignancies, such as the liver, peritoneum, abdominal wall, bowel surface, and other intraperitoneal organs and surfaces. A total of 576 electrocautery markings (288 single and 288 double lesions) were made. Detection of these lesions was considered the main outcome variable of the study. The time allowed for this test was a maximum of 20 minutes.

Zoom
Fig. 3 Study design for comparison, in a porcine model, of peritoneoscopy via the transgastric or transcolonic route and transabdominal laparoscopy using a rigid or a flexible laparoscope. Each animal underwent examination by two methods that were each performed by a different examiner (operator 1 and operator 2). With these test sequences, each animal was able to undergo double testing while the observers were kept blinded to the placement of the electrocautery markings. VAS, visual analogue scale; TAR, transabdominal peritoneoscopy using a standard rigid laparoscope; TAF, transabdominal peritoneoscopy using a flexible endoscope; TC, transcolonic peritoneoscopy; TG, transgastric peritoneoscopy.


Sequence of tests and study procedure

Three investigators (D.v.R., A.S., K.C.), all with extensive experience in flexible endoscopy, diagnostic laparoscopy, and NOTES, took part in the study to compare the four approaches (TAR, TAF, TG, TC) in a random order. Each of the four methods was tested in relation to the three main study parameters (visualization of organs, biopsy targeting simulation, detection of premarked lesions), as described in detail below. For ethical reasons, each animal underwent examination by two different methods each performed by a different examiner (operator 1 and operator 2 in the flow-chart), using the sequence shown in the flow-chart ([Fig. 3]). Using this test sequence, each animal was able to undergo double testing and observers were kept blinded to the placement of the electrocautery markings, as they were not allowed to enter the operating room during these procedures.


Description of the four approaches

(a) Transabdominal laparoscopy using a standard rigid laparoscope (TAR)

Systematic peritoneoscopy was performed using the 5.5-mm umbilical trocar and a 5-mm, 30°-angled viewing video laparoscope (EndoEYE; Olympus) with an Exera II CLV-180 light source and Exera II CV-180 processor (Olympus). The animal was placed in the Trendelenburg, anti-Trendelenburg, 30° left and 30° right positions in a systematic fashion to facilitate optimal examination of all intraperitoneal organs and areas. Additional positioning was facilitated as requested by the observer. The amount of time taken for the examination was noted. The quadrant location of each identified electrocautery marking was recorded.


(b) Transabdominal laparoscopy using a flexible endoscope (TAF)

Systematic peritoneoscopy was performed using the 5.5-mm umbilical trocar, a flexible 5-mm video endoscope (GIF N180) with the Exera II CLV-180 light source and processor. Positioning of the animal, systematic evaluation, and time allowed for evaluation and documentation of each identified electrocautery marking were all carried out in the same way as described for TAR.


(c) Transgastric peritoneoscopy using a flexible endoscope (TG)

TG was performed using a standard two-channel therapeutic upper gastrointestinal endoscope (GIF 2T160; Olympus) with the Exera II CLV-180 light source and processor. Transgastric access to the peritoneal cavity was obtained using the transoral route and the percutaneous endoscopic gastrostomy (PEG) technique. This approach was chosen because it has demonstrated an improved safety profile compared with other transgastric access methods in previous studies [29] [30]. External abdominal pressure dimpling was used to locate the anterior gastric wall. Gastric access was created by needle-knife puncture. A 0.035-inch guide wire (Jagwire; Boston Scientific Corporation, Natick, Massachusetts, USA) was passed through the needle, and the needle was then exchanged for an 18-mm dilating balloon (CRE dilation balloon; Boston Scientific Corporation). The balloon was used to dilate the tract and gain entry to the abdominal cavity. Positioning of the animal, systematic evaluation, and time allowed for evaluation and documentation of each identified electrocautery marking were as previously described for TAR.


(d) Transcolonic peritoneoscopy using a flexible endoscope (TC)

TC was performed using the GIF 2T160 with the Exera II CLV-180 light source and processor. Transcolonic access to the peritoneal cavity was obtained using the transanal route and needle-knife puncture of the sigmoid colon at a distance of about 15 – 20 cm from the anus. A 0.035-inch guide wire was passed through the needle-knife incision and an 18-mm dilating balloon was used to dilate the incision and gain entry to the peritoneal cavity. Positioning of the animal, systematic evaluation, and time allowed for evaluation and documentation of each identified electrocautery marking were as previously described for TAR.



Outcomes and statistical analysis

The primary outcome variable was the number of electrocautery markings found during laparoscopy/peritoneoscopy, meaning the rate of detection of the premarked lesions.

Secondary outcomes variables were:

  • VAS score for organ visualization, on a scale of 0 – 10

  • Biopsy targeting simulation, meaning the ability to touch organs with the instrument tip

  • Procedure-related complications, such as bleeding, organ or vessel injury, perforations, and cardiopulmonary complications during peritoneoscopy

For the primary outcome, a one-tailed noninferiority design was used to compare TAR with TG. The published literature shows a mean sensitivity of 79.1 % for TAR for metastatic disease in upper gastrointestinal, hepatic, and biliopancreatic malignancies [2] [5] [31]. Using α = 0.025 and 1−β = 0.8, assuming TAR has a sensitivity of at least 75 %, and rejecting alternative approaches as inferior at a sensitivity of at least 65 % (margin of equivalence of 10 %), it was determined that a sample size of at least 142 electrocautery markings for each group was required. To ensure the electrocautery markings were evenly distributed between the pigs, a total of 144 (12 × 12) electrocautery markings were made for each group.

Statistical calculations were carried out using SPSS 14.0 (SPSS Inc., Chicago, Illinois, USA). Quantitative data are expressed as means and confidence intervals. The diagnostic yields for the detection of electrocautery markings and organ surface visualization were compared with a two-tailed Mann – Whitney U test, and the ability to touch the organ or target area with the instrument tip was compared using Fisher’s exact test. P < 0.05 was considered significant. In order to carry out these statistical techniques, it was assumed that all observations were statistically independent, even though more than one observation originated from each individual pig.



Results

Detection rate of intraperitoneal lesions

The sensitivities for the detection of simulated metastases (the main outcome variable) are shown in [Fig. 4] and were 78.5 %, 59.7 %, 48.6 %, and 38.9 % for TAR, TAF, TC, and TG, respectively. Standard laparoscopy (TAR) was superior to all other techniques using a flexible endoscope, with highly significant differences (P < 0.01). TAF was also superior to TC (P = 0.03) and TG (P < 0.01). Overall, there were no statistically significant differences between the two NOTES approaches (TC vs. TG, P = 0.11). However, the transcolonic approach was superior for examining the upper abdomen in comparison with the transgastric approach (P = 0.03). More detailed data on sensitivity rates for lesion detection in specific intra-abdominal areas are provided in [Fig. 5], [Fig. 6], [Fig. 7], [Fig. 8], [Fig. 9], [Fig. 10].

Zoom
Fig. 4 Sensitivity for lesion detection in the peritoneal cavity. Values shown are the percentage of lesions identified in each group (n = 144 lesions per group); error bars indicate 95 % confidence intervals. TAR, transabdominal peritoneoscopy using a standard rigid laparoscope; TAF, transabdominal peritoneoscopy using a flexible endoscope; TC, transcolonic peritoneoscopy; TG, transgastric peritoneoscopy.
Zoom
Fig. 5 Sensitivity for the detection of lesions in the upper abdomen. Values shown are the percentage of lesions identified in each group (n = 72 lesions per group); error bars indicate 95 % confidence intervals. TAR, transabdominal peritoneoscopy using a standard rigid laparoscope; TAF, transabdominal peritoneoscopy using a flexible endoscope; TC, transcolonic peritoneoscopy; TG, transgastric peritoneoscopy.
Zoom
Fig. 6 Sensitivity for the detection of lesions in the lower abdomen. Values shown are the percentage of lesions identified in each group (n = 72 lesions per group); error bars indicate 95 % confidence intervals. TAR, transabdominal peritoneoscopy using a standard rigid laparoscope; TAF, transabdominal peritoneoscopy using a flexible endoscope; TC, transcolonic peritoneoscopy; TG, transgastric peritoneoscopy.
Zoom
Fig. 7 Sensitivity for the detection of lesions in the right upper abdominal quadrant. Values shown are the percentage of lesions identified in each group (n = 36 lesions per group); error bars indicate 95 % confidence intervals. TAR, transabdominal peritoneoscopy using a standard rigid laparoscope; TAF, transabdominal peritoneoscopy using a flexible endoscope; TC, transcolonic peritoneoscopy; TG, transgastric peritoneoscopy.
Zoom
Fig. 8 Sensitivity for the detection of lesions in the left upper abdominal quadrant. Values shown are the percentage of lesions identified in each group (n = 36 lesions per group); error bars indicate 95 % confidence intervals. TAR, transabdominal peritoneoscopy using a standard rigid laparoscope; TAF, transabdominal peritoneoscopy using a flexible endoscope; TC, transcolonic peritoneoscopy; TG, transgastric peritoneoscopy.
Zoom
Fig. 9 Sensitivity for the detection of lesions in the right lower abdominal quadrant. Values shown are the percentage of lesions identified in each group (n = 36 lesions per group); error bars indicate 95 % confidence intervals. TAR, transabdominal peritoneoscopy using a standard rigid laparoscope; TAF, transabdominal peritoneoscopy using a flexible endoscope; TC, transcolonic peritoneoscopy; TG, transgastric peritoneoscopy.
Zoom
Fig. 10 Sensitivity for the detection of lesions in the left lower abdominal quadrant. Values shown are the percentage of lesions identified in each group (n = 36 lesions per group); error bars indicate 95 % confidence intervals. TAR, transabdominal peritoneoscopy using a standard rigid laparoscope; TAF, transabdominal peritoneoscopy using a flexible endoscope; TC, transcolonic peritoneoscopy; TG, transgastric peritoneoscopy.

Visualization of organs

The VAS scores for organ surface visualization are shown in [Fig. 11] and were 7.15, 6.71, 5.07, and 4.35 for TAR, TAF, TC, and TG, respectively. Standard laparoscopy (TAR) was again significantly superior to both NOTES approaches (P < 0.01). TAR was superior in comparison with TAF (P = 0.04). The transcolonic approach was superior to the transgastric approach (TC vs. TG, P < 0.01).

Zoom
Fig. 11 Visual analogue scale (VAS) assessment of organ surface visualization. Values shown are means; error bars indicate 95 % confidence intervals. TAR, transabdominal peritoneoscopy using a standard rigid laparoscope; TAF, transabdominal peritoneoscopy using a flexible endoscope; TC, transcolonic peritoneoscopy; TG, transgastric peritoneoscopy.

Biopsy targeting simulation

The success rates (percentages of all attempts) for touching organ surfaces with the instrument tip and simulating biopsies are shown in  [Fig.12] and were 87 %, 85 %, 72 %, and 65 % for TAR, TAF, TC, and TG, respectively. Standard laparoscopy was significantly superior to both NOTES approaches (TAR vs. TC, P = 0.03; TAR vs. TG, P < 0.01). There were no statistically significant differences between the two transabdominal approaches (TAR vs. TAF, P = 0.8). There were no statistically significant differences between the two NOTES approaches (TC vs. TG, P = 0.28).

Zoom
Fig. 12 Success rate for simulation of organ biopsy. Values shown are the percentage of successful attempts to touch the surface of an organ with the tip of the endoscope; error bars indicate 95 % confidence intervals. TAR, transabdominal peritoneoscopy using a standard rigid laparoscope; TAF, transabdominal peritoneoscopy using a flexible endoscope; TC, transcolonic peritoneoscopy; TG, transgastric peritoneoscopy.

Procedure-related complications

No procedure-related complications occurred during placement of the abdominal trocars. Peritoneal access was achieved without complications in all TG and TC cases. No other complications occurred during any of the laparoscopies in these acute experiments.



Discussion

NOTES represents a major paradigm shift toward scarless endoscopic surgery, and the adoption of NOTES peritoneoscopy into clinical practice has recently begun. The benchmark for any NOTES procedure will be the safety and efficacy of the corresponding laparoscopic procedure, as these are minimally invasive, highly effective, and have low complication rates. Thorough preclinical investigation in randomized controlled trials to evaluate the efficacy and safety of each proposed NOTES indication is required in order to avoid premature clinical implementation that could potentially endanger patients.

The first experimental application of NOTES consisted of diagnostic laparoscopy through a transgastric approach [24]. Another recently published animal trial compared standard peritoneoscopy with the transgastric NOTES approach in the detection of intraperitoneal color-coded beads [32] The study compared only two modalities (TAR and TG) and used only a total of 64 intraperitoneal color-coded beads to simulate metastases. The authors reported a sensitivity for transabdominal laparoscopy of 95 %, which is higher than the sensitivity found in the present study. This might be due to the fact that color-coded beads are easier to detect than electrocautery lesions. Interestingly, the reduction in sensitivity that resulted from the transgastric NOTES approach was 32 percentage points in the Dutch study [32], which is quite similar to the present data.

In comparison with that study, the design of the present study was intended to minimize any potential bias in lesion detection by using a flexible endoscope for placement and using the lateral ports strictly for placement of the electrocautery markings only. However, both of these studies cast substantial doubt on the clinical value of NOTES peritoneoscopy in its present form for diagnostic purposes, such as tumor staging. Despite this, human use of transluminal peritoneoscopy has already been reported even before systematic assessment of the method in animal trials [27] [28].

The present study shows that NOTES peritoneoscopy is inferior to standard rigid laparoscopy in every respect tested. This was due to the more straightforward access through the abdominal skin, as well as the use of a rigid rather than a flexible instrument. When the flexible scope was used for transcutaneous laparoscopy, it was equivalent to the rigid instrument for organ visualization and biopsy purposes, but was inferior to it for the detection of lesions in the peritoneum. The inferiority of transluminal laparoscopy can therefore only partly be explained by the use of flexible scopes, with their limited access.

The original idea of achieving easy and effective transgastric laparoscopy will therefore probably have to be discarded. In this study, the transgastric NOTES approach proved to be inferior to all other modalities and yielded a sensitivity of only 38.9 % – a decrease in sensitivity of 40 % compared with standard laparoscopy. Transgastric peritoneoscopy was not an adequate procedure, particularly for the detection of lesions in the upper abdomen. This is partly explained by the reduced stability and reduced anatomic orientation due to the retroflexion of the endoscope. An additional factor is that about 50 – 60 cm of the endoscope is needed to reach the gastrotomy. This further limits the length of the endoscope and its maneuverability in the peritoneal cavity and explains the inferior outcome in comparison with transrectal peritoneoscopy. A recent study has also indicated that a trans-sigmoid route is superior to the transgastric route for access to the upper abdomen [33]. Transgastric NOTES peritoneoscopy produced the poorest results for the detection of lesions in the upper right abdomen (27.7 % sensitivity, compared with 83.3 % sensitivity for standard laparoscopy). This is especially worrisome because the detection of metastatic lesions in the liver is one of the main purposes of diagnostic laparoscopy.

The model used to simulate peritoneal metastases yielded a sensitivity of 78.5 % for standard rigid laparoscopy. This corresponds well to clinical data on the sensitivity of diagnostic laparoscopy. The literature for esophageal cancer shows sensitivities of 71 % for finding peritoneal metastases, 78 % for finding nodal metastases, and 86 % for finding liver metastases [2]. For gastric cancer, diagnostic laparoscopy has a sensitivity of 69 % for detecting peritoneal metastases and 96 % for detecting hepatic metastases [31]. For pancreatic adenocarcinoma, it has a median sensitivity of 94 %, and for biliary cancer, a sensitivity of 60 % [5].

The method used to inspect the peritoneal cavity was identical to the standard minilaparoscopy technique with a single paraumbilical port and without inspection of the omental bursa; in Germany, this technique is routinely performed by gastroenterologists for cancer staging [34] [35] [36] [37]. Minilaparoscopy has been shown to have a comparable diagnostic sensitivity for inspection of the peritoneal cavity and preoperative cancer staging, with optimized cosmetic results in comparison with extended diagnostic laparoscopy [34] [35] [36] [37]. All of the investigators who participated in the present trial routinely perform this kind of diagnostic laparoscopy in their clinical practice. Interestingly, they also all have a higher caseload for flexible endoscopy in their clinical practice than for diagnostic laparoscopy. Moreover, all of the investigators have extensive experience with more than 100 flexible NOTES procedures in the animal model. One would expect this study setting to favor the flexible endoscopic approach. Nevertheless, the outcome of the study shows that even in this setting, standard laparoscopy was significantly superior to all of the flexible endoscopic approaches.

When the study was being designed, it was expected that the flexible endoscope would provide access to areas that the rigid instrument could not easily reach. In contrast to these expectations, the results show that using a flexible instrument in the peritoneal cavity can result in a loss of spatial orientation and associated problems during a structured examination. For example, during attempts to access a peritoneal area such as the upper left quadrant, even a slight deflection of the endoscopic tip can result in loss of spatial orientation, so that the endoscope ends up in areas other than those initially intended. The gastrointestinal tract provides navigation structures for the flexible endoscope in the form of the gut lumen, and it seems that guiding structures of this type are needed with the flexible instrument to ensure that both stability and spatial orientation are maintained.

The possibility that operators with a different training background (gastroenterology versus surgery) might produce different results in experimental or clinical performance should be considered. However, the Dutch study and one recent study that compared trans-sigmoid with transgastric peritoneoscopy demonstrated similar results to our study, with TAR being superior to TG, and TC being superior to TG [32] [33] [38]. These results were achieved in a setting where surgeons were performing TAR and a group comprised of gastroenterologists and surgeons were comparing TC with TG [32] [33] [38]. Therefore, further improvements in the endoscopic platforms available for NOTES peritoneoscopy appear to be necessary before it can be recommended for use in clinical trials.

Some limitations of this study and the animal model also need to be considered. Firstly, there are some differences between porcine and human anatomy. The porcine liver has a larger middle lobe, which can sometimes partially cover the other lobes. For evaluation, the middle and the right lobe were regarded as representing the human right lobe. In addition, the porcine uterus has two uterine horns instead of the single uterine body in humans, and the porcine spleen tends to be longer. Despite these anatomical differences, there are multiple similarities with regard to size, anatomy, and organ location. The present study was designed to compare the ability of different diagnostic modalities to navigate and to detect subtle lesions in a similar environment. The porcine model appears to be ideal for a comparison of standard laparoscopy with flexible endoscopic approaches in an adequately powered study in controlled conditions.

Another limitation of the study was that organ visualization using a VAS might be regarded as a rather subjective outcome variable, particularly as it was done by the examiners, rather than by an independent person who would preferably be blinded to the access method. This parameter was therefore only treated as a secondary outcome. It is nevertheless helpful to evaluate the parameter, as it makes it possible to evaluate spatial orientation and also reflects the ease and comfort of diagnostic peritoneoscopy for the investigator. The outcomes also correlated with the lesion detection rate.

A third limitation of the trial is that the currently used endoscopic optics are designed and optimized for use in the gastrointestinal tract. Therefore, the use of commercially available flexible endoscopes might be a limiting factor for NOTES, because in contrast, the current laparoscopic optics and instruments are optimized for intraperitoneal use. Future developments involving improved optics and specialized NOTES platforms and instruments may therefore need to be evaluated. Moreover, this study only compares the NOTES approach with laparoscopy for its diagnostic efficiency. There may still be several therapeutic interventions, such as biopsy of the liver, where the performance of NOTES is equivalent or superior to its laparoscopic counterpart.

In conclusion, this prospective, blinded, randomized controlled trial showed that all the flexible endoscopic approaches were inferior to standard diagnostic laparoscopy with rigid instruments. In its current form, NOTES appears to be unsuitable for diagnostic laparoscopy.



Competing interests: None.

Acknowledgments

The authors would like to thank Olympus Germany for providing material support (endoscopic and laparoscopic equipment) and partially funding the animal procedures for the study. We would also like to thank Bettina Riecken, MD MPH, and Heiko Pohl, MD MPH, for their advice on statistical issues.


Corresponding author

D. von Renteln, MD
Department of Interdisciplinary Endoscopy
University Hospital Hamburg-Eppendorf
Martinistr. 52
20246 Hamburg
Germany   
Fax: +49-040-741040004   


Zoom
Fig. 1 Port location for creation of lesions within the peritoneal cavity and examination by two of the four approaches. a The locations at which the one paraumbilical trocar and four lateral trocars were introduced into the anesthetized pigs. b Photograph of two of the lateral 5.5-mm trocars that were used to create the electrocautery markings positioned ready for use. c Photograph of transabdominal diagnostic peritoneoscopy with a flexible endoscope being performed via the single paraumbilical trocar.
Zoom
Fig. 2  Example lesions in the model of peritoneal metastases. a Laparoscopic view of a single electrocautery marking on the gallbladder. b Endoscopic view of a double electrocautery marking on the gallbladder. c Laparoscopic view of a single electrocautery marking on the liver. d Endoscopic view of a single electrocautery marking on the liver. e Laparoscopic view of a single electrocautery marking on the abdominal wall. f Endoscopic view of a double electrocautery marking on the abdominal wall.
Zoom
Fig. 3 Study design for comparison, in a porcine model, of peritoneoscopy via the transgastric or transcolonic route and transabdominal laparoscopy using a rigid or a flexible laparoscope. Each animal underwent examination by two methods that were each performed by a different examiner (operator 1 and operator 2). With these test sequences, each animal was able to undergo double testing while the observers were kept blinded to the placement of the electrocautery markings. VAS, visual analogue scale; TAR, transabdominal peritoneoscopy using a standard rigid laparoscope; TAF, transabdominal peritoneoscopy using a flexible endoscope; TC, transcolonic peritoneoscopy; TG, transgastric peritoneoscopy.
Zoom
Fig. 4 Sensitivity for lesion detection in the peritoneal cavity. Values shown are the percentage of lesions identified in each group (n = 144 lesions per group); error bars indicate 95 % confidence intervals. TAR, transabdominal peritoneoscopy using a standard rigid laparoscope; TAF, transabdominal peritoneoscopy using a flexible endoscope; TC, transcolonic peritoneoscopy; TG, transgastric peritoneoscopy.
Zoom
Fig. 5 Sensitivity for the detection of lesions in the upper abdomen. Values shown are the percentage of lesions identified in each group (n = 72 lesions per group); error bars indicate 95 % confidence intervals. TAR, transabdominal peritoneoscopy using a standard rigid laparoscope; TAF, transabdominal peritoneoscopy using a flexible endoscope; TC, transcolonic peritoneoscopy; TG, transgastric peritoneoscopy.
Zoom
Fig. 6 Sensitivity for the detection of lesions in the lower abdomen. Values shown are the percentage of lesions identified in each group (n = 72 lesions per group); error bars indicate 95 % confidence intervals. TAR, transabdominal peritoneoscopy using a standard rigid laparoscope; TAF, transabdominal peritoneoscopy using a flexible endoscope; TC, transcolonic peritoneoscopy; TG, transgastric peritoneoscopy.
Zoom
Fig. 7 Sensitivity for the detection of lesions in the right upper abdominal quadrant. Values shown are the percentage of lesions identified in each group (n = 36 lesions per group); error bars indicate 95 % confidence intervals. TAR, transabdominal peritoneoscopy using a standard rigid laparoscope; TAF, transabdominal peritoneoscopy using a flexible endoscope; TC, transcolonic peritoneoscopy; TG, transgastric peritoneoscopy.
Zoom
Fig. 8 Sensitivity for the detection of lesions in the left upper abdominal quadrant. Values shown are the percentage of lesions identified in each group (n = 36 lesions per group); error bars indicate 95 % confidence intervals. TAR, transabdominal peritoneoscopy using a standard rigid laparoscope; TAF, transabdominal peritoneoscopy using a flexible endoscope; TC, transcolonic peritoneoscopy; TG, transgastric peritoneoscopy.
Zoom
Fig. 9 Sensitivity for the detection of lesions in the right lower abdominal quadrant. Values shown are the percentage of lesions identified in each group (n = 36 lesions per group); error bars indicate 95 % confidence intervals. TAR, transabdominal peritoneoscopy using a standard rigid laparoscope; TAF, transabdominal peritoneoscopy using a flexible endoscope; TC, transcolonic peritoneoscopy; TG, transgastric peritoneoscopy.
Zoom
Fig. 10 Sensitivity for the detection of lesions in the left lower abdominal quadrant. Values shown are the percentage of lesions identified in each group (n = 36 lesions per group); error bars indicate 95 % confidence intervals. TAR, transabdominal peritoneoscopy using a standard rigid laparoscope; TAF, transabdominal peritoneoscopy using a flexible endoscope; TC, transcolonic peritoneoscopy; TG, transgastric peritoneoscopy.
Zoom
Fig. 11 Visual analogue scale (VAS) assessment of organ surface visualization. Values shown are means; error bars indicate 95 % confidence intervals. TAR, transabdominal peritoneoscopy using a standard rigid laparoscope; TAF, transabdominal peritoneoscopy using a flexible endoscope; TC, transcolonic peritoneoscopy; TG, transgastric peritoneoscopy.
Zoom
Fig. 12 Success rate for simulation of organ biopsy. Values shown are the percentage of successful attempts to touch the surface of an organ with the tip of the endoscope; error bars indicate 95 % confidence intervals. TAR, transabdominal peritoneoscopy using a standard rigid laparoscope; TAF, transabdominal peritoneoscopy using a flexible endoscope; TC, transcolonic peritoneoscopy; TG, transgastric peritoneoscopy.