scarce interest of centres

Objective: Recent non-randomized studies suggest that extended endoscopic submucosal dissection (ESD) is equally effective in removing large rectal adenomas as transanal endoscopic microsurgery (TEM). If equally effective, ESD might be a more cost-effective approach as this strategy does not require expensive equipment, general anesthesia and hospital admission. Furthermore, ESD appears to be associated with fewer complications. In a randomized trial we will compare the cost-effectiveness and cost-utility of TEM and ESD for the resection of large rectal adenomas.

Study design: 15 centers will participate in this multicenter randomized trial comparing TEM versus ESD.

Study population: Patients with a large rectal adenoma (≥2cm), located between 2 and 15 cm from the anal verge. Invasive cancer is excluded by histopathology and endoscopic ultrasonography. Patients must be in a health condition that permits general anesthesia.

Interventions: Patients will be randomized between

a. TEM: under general anesthesia b. ESD under sedation

1. a TEM tube will be inserted in the rectum. With specialized instruments the adenoma will be dissected en bloc by a full thickness excision, after which the patient will be admitted to the hospital.
2. an endoscope will be inserted into the rectum and the submucosa underneath the lesion will be injected with saline to lift the adenoma. With an endoscopic knife (Insulated Tip Knife, Olympus or Water Jet, Erbe) the lesion will be resected through the submucosal plane in an eb-bloc fashion, after which the patient will be observed for at least 24h in-hospital.

Primary Endpoint: incidence of recurrence at 12 months

Secondary Endpoints:

morbidity, subdivided into major (requiring surgery) and minor (requiring endoscopic or medical intervention) anorectal function. disease specific and general quality of life; number of days not spent in hospital from initial treatment until 2 years afterwards; adenoma

Sample size: Assuming a comparable baseline recurrence rate for TEM and ESD of 6% and considering an upper limit of 10% for ESD to be non-inferior (beta-error 0.2 and one-sided alpha-error 0.05), 60 patients are needed per group. These numbers provide sufficient power to reveal relevant differences in expected morbidity and in number of days not spent in hospital.

Economic evaluation: A cost-effectiveness and cost-utility analysis of ESD against TEM for large rectal adenomas from a societal perspective with respectively the costs per recurrence free patient and the cost per quality adjusted life year as primary outcome measures.

PROBLEM DEFINITION Rectal cancer is a common disease in Western countries, increased with high age, male sex and obesity (1,2). As for other districts, premalignant intraepithelial neoplasia inside a rectal adenoma precedes the occurrence of invasive rectal cancer (3,4). Early endoscopic detection and removal of rectal adenomas prevents the development of rectal cancer and is therefore the most reliable contributor to the 'cure' of this disease (5,6). When rectal adenomas become large, however, standard endoscopic therapies like loop polypectomy or one-step endoscopic resection result inadequate. Therefore, large rectal adenomas must be removed en-bloc either surgically or by extended endoscopic submucosal dissection (ESD).

In 1983 a novel surgical approach for the resection of large rectal adenomas has been introduced in the clinical practice in Germany: Transanal Endoscopic Microsurgery (TEM).(7) This procedure encompasses general anesthesia and the use of expensive specialized equipment. On the other hand, it generally allows a full-thickness rectal wall excision. Since its introduction, many surgical practices have adopted TEM as the new standard therapy for large rectal adenomas (8). In more recent years advanced endoscopic therapies like extended ESD have rapidly evolved.(9).

For extended ESD an en-bloc specimen, including mucosa and consistent portion of the submucosal layer, are resected instead of the full-thickness rectal wall, combining the advantages of an en-bloc resection with potential benefit of fewer complications.

Supporters of the TEM technique praise the excellent exposure of the rectum and the minimal invasiveness, as opposed to conventional surgical techniques.(10-12) Besides, recurrence rates after TEM appear to be lower when compared to conventional surgical transanal excision.(13) The TEM technique has shown to be highly efficacious in several retrospective and prospective case series with reported recurrence rates of 0-19% and complication rates of 2-21%.(14-23) On the other hand, extended ESD has gained more and more support in the last few years, mainly due to good clinical results after ESD of neoplasia in the esophagus and stomach as reported in Japanese centers (24,25) ESD has also been described for the treatment of large colorectal adenomas, revealing recurrence rates of 0-9% and complication rates of only 0-9% (9,27). If adenomas could not be removed completely during one ESD attempt, repeat ESD for residual disease generally led to an overall success rate of 96-100%. In general, all recurrences were detected during the first control endoscopy after 3 months; repeat endoscopic resection of residual disease led to an overall success rate of almost 100%.

Since the efficacy of extended ESD for large rectal adenomas appears to be comparable to TEM, we decided to design this randomized trial. In fact, until now, TEM and ESD have never been formally compared, and no such comparative studies have been registered at this moment. Although selection bias inevitably exists in prospective and retrospective case series, the results of these studies suggest that both TEM and ESD have comparable recurrence rates. Even when recurrences occur after TEM or ESD, most of these can successfully be re-treated without the need for radical surgery. The literature furthermore suggests that ESD is associated with fewer complications, reduced hospital admission, and no general anesthesia is required for ESD, all of which are favorable in both patients' and societal perspective. These contrasts of the two procedures might well lead to differences in costs and quality of life. Therefore, we designed a multicentre randomized trial to compare TEM and ESD for the resection of large rectal adenomas.

RELEVANCE Colorectal cancer (CRC) is the second most common cancer in Western countries. Rectal cancer accounts for approximately 40% of CRC cases. The treatment of rectal cancer encompasses a multidisciplinary collaboration including gastroenterologists, surgeons, oncologists, radiotherapists and specialized nurse practitioners. Standard therapy consists of radical surgery in combination with radiotherapy (and possibly chemotherapy), which have major morbidity and mortality (28). Therefore, this disease has a major impact on health care services (29).

Since early detection and removal of rectal adenomas prevents the occurrence of rectal cancer, CRC screening has been adopted in many western countries (30). When CRC screening was introduced, this inevitably lead to an increased detection of early rectal neoplasia (31). It is therefore expected that more rectal adenomas will need endoscopic or surgical treatment in the forthcoming years. Consequently, the most appropriate therapy concerning efficacy, safety, quality of life and costs must be selected to deal with the expected increase in rectal adenomas.

Traditionally, adenomatous colorectal lesions that could not be resected endoscopically were referred for surgery. Conventional surgical approaches like radical surgery and trans-sphincteric or trans-sacral operations have nowadays been replaced by TEM, since this procedure has higher efficacy and lower morbidity. In recent years endoscopic therapies have further evolved, as a result of which large rectal adenomas more often are treated by flexible endoscopy. In case series, endoscopic resection of large colorectal adenomas has led to recurrence rates that were slightly higher than to TEM (18.0% versus 6%), but all retreated endoscopically, and to complication rates that appeared to be low and equivalent if not slightly lower (4% versus 8%) (32). Furthermore, ESD can safely be performed under conscious sedation, and short hospital admission is required as opposed to TEM. The reduced morbidity, reduced hospital admission and redundancy of anesthesia associated with ESD are beneficial from both patients' and societal perspective.

OBJECTIVE Hypothesis Transanal endoscopic microsurgery (TEM) and extended endoscopic submucosal dissection (ESD) are both effective treatments for large rectal adenomas with comparable recurrence rates. However, ESD does not require general/spinal anesthesia and may be associated with lower morbidity. Therefore, ESD may improve quality of life and reduce health care costs.

Objective The main objective of the proposed randomized study is to compare the effectiveness of TEM and ESD for the removal of large rectal adenomas.

Primary Endpoint: incidence of recurrence at 3 months

Secondary Endpoints:

morbidity, subdivided into major (requiring surgery) and minor (requiring endoscopic or medical intervention) anorectal function. disease specific and general quality of life; number of days not spent in hospital from initial treatment until 12 months

STRATEGY Design This will be a multicentre randomized trial comparing TEM and ESD in patients with large rectal adenomas with respect to cost-effectiveness and safety.

Randomization Patient data are entered into a computerized database and by means of an unchangeable computer generated number patients will be randomized to undergo TEM or ESD.

Blinding Blinding of patients and physicians during treatment is unfeasible, since the two treatment strategies are highly different in nature and in associated care.

Study population

Patients are eligible for this trial when they meet the following inclusion criteria:

Diagnosed with a large non-pedunculated rectal adenoma (sessile or flat) with a largest diameter of ≥2 cm (estimated by an opened resection snare of 20 or 30 mm).

The lower and upper borders of the adenoma are located at ≥2 cm and ≤15 cm from the anal verge, respectively.

Biopsies of the lesion did not show malignant neoplastic tissue on histopathological evaluation; only lesions with low or high grade dysplasia are suitable for inclusion.

During flexible video endoscopy there are no signs of endoscopic suspicion for submucosal invasive cancer (Kudo pit pattern type V; excavated/depressed type morphology; fold convergence; or large smooth nodule >1 cm in a flat lesion) (33). In case of doubt, patients will undergo EUS as described at (2).

In case doubt remains after flexible video endoscopy, endoscopic ultrasonography (EUS) of the rectal adenoma should exclude invasion into the submucosal layer and exclude pathological lymphadenopathy (lymph nodes >1 cm). When pathological lymph nodes are present, fine needle aspiration will be performed to exclude lymph node metastasis (N+ disease).

If not performed already, total colonoscopy will be done to detect and remove all synchronous colonic adenomas or cancers first. Cecal intubation must be confirmed by identification of the appendiceal orifice and ileocecal valve.

The general health condition of the patient permits general anesthesia (ASA- classification I-III).

Absence of non-correctable coagulopathy (international normalized ratio >1,5, or platelet count <90 × 109/l).

Patient age of 18 years or older.

Intervention strategies Transanal endoscopic microsurgery: TEM will be performed as described by Buess.(7) Under general anesthesia a specialized TEM rectoscope of 12 or 20 cm in length (Wolf GmbH Knittlingen) or 15 cm in length (Storz GmbH Tuttlingen, Germany) is inserted within the rectum to assure proper visualization of the lesion. The rectoscope is fixed to the operating table by a supporting device, providing the opportunity to reposition the rectoscope during ongoing surgery. The rectal cavity is insufflated with CO2 by a combined endosurgical unit to achieve constant distension for appropriate visualization of the rectal adenoma. The combined endosurgical unit further regulates irrigation and suction, thereby maintaining a constant intra rectal pressure. With the use of a binocular stereoscopic eyepiece for three-dimensional view (Wolf GmbH only) or a forward oblique telescope (Storz GmbH) a magnified view is being created for visualization of the lesion. With various HF monopolar and bipolar instruments, Ultracision harmonic scalpel (Ethicon Endo-Surgery, Cincinnati, USA), needle diathermy, tissue handling forceps, needle holder, suction probe, injection needle, clip applicator) the adenoma will be dissected by means of an en-bloc full-thickness rectal wall excision until the perirectal fat. Postoperatively, patients will preserve a urinary catheter that will be removed at the first postoperative day. Patients are admitted to the hospital in accordance with current practice.

After 3 months a control flexible endoscopy will be performed by the surgeon who performed the TEM procedure. If presumed residual disease is seen, biopsies will be taken to confirm the presence of neoplasia by histology. Hereafter, the residual adenomatous tissue will be resected endoscopically by either ESD (if >20 mm) or endoscopic mucosal resection (EMR) (if <20mm). Any intervention by ESD/EMR at 3 months is part of the TEM treatment strategy.

Endoscopic submucosal dissection: Endoscopic submucosal dissection is performed as described by Fujiro (9) At the discretion of the endoscopist, conscious sedation is used with 2.5-10 mg midazolam and/or 25-100 µg of fentanyl. An endoscope is inserted into the rectum and air insufflation via the endoscope will provide proper distension of the rectum. The submucosa underneath the lesion will be injected through an endoscopic injection catheter with a solution of saline 0.9%, 1 ml methylene blue, and 1:10,000 units adrenaline in order to lift the adenoma (no upper volume limit). The lesion will be marked around with a 5 mm disease free margin by APC dots. An endoscopic knife (Insulated Tip Knife, Olympus or Water Jet, Erbe) will be used to resect the lesion through the submucosal plane in an en-bloc fashion. Visible submucosal vessels will be treated by endoscopic clips or electro-coagulation to prevent delayed bleeding. Hereafter, the edges of the mucosal defect and potential remnants within the resection crater will always be treated with APC to increase adenoma clearance. If bleeding during the procedure precludes >90% clearance of the adenoma, the ESD procedure will be considered as failure. All patients are admitted to the hospital for at least 24-hours observation; in accordance with current practice.

After 3 months the treating endoscopist will perform a control endoscopy. If presumed residual disease is seen, biopsies will be taken to confirm the presence of neoplasia by histology. Hereafter, the residual adenomatous tissue will be resected endoscopically by either ESD (if >20 mm) or endoscopic mucosal resection (EMR) (if <20mm). Any intervention by ESD/EMR at 3 months is part of the TEM treatment strategy.

Intervention failure When for technical reasons ESD procedures turn out not to be performable after randomization or when ESD fails to remove >90% of the adenoma, the patient will automatically undergo the TEM treatment strategy.

Histopathological evaluation Resection specimens after TEM and ESD will be stretched and pinned on a cork plate before immersion into formalin. After standard processing the resection specimen will be transected each cm for evaluation by a gastrointestinal pathologist. The lateral and basal resection margins will be evaluated for absence of neoplasia, when possible.

The risk of lymph node metastases is increased in case of malignant neoplasia extending into the submucosal layer, poor tumor differentiation, mucinous cancer, vascular invasion and tumor budding, all of which warrant further radical surgery (34-37). By strict adherence to the inclusion criteria, the risk of invasive cancer is reduced to 1.6-3% (33,38). In case of an unexpected invasive cancer despite adherence to the inclusion criteria, the patient will be excluded after the histopathological evaluation (late exclusion). In case of intramucosal cancer (i.e. not invading through the muscularis mucosae), both TEM and ESD will be regarded as sufficient treatment when the lesion is radically removed.

Outcome parameters

Primary outcome measure (for non-inferiority):

Recurrence of neoplasia, defined as the presence of histologically proven neoplastic tissue in either visible recurrent lesions or in random biopsies, taken at surveillance endoscopies after the intervention strategy has been completed.

Any remnant adenoma identified and treated by ESD/EMR/APC at 3 months is considered part of the initial intervention strategy in both arms. Hereafter, patients will undergo surveillance endoscopies at 6 and 12 months by an independent endoscopist who is blinded for the primary treatment. During each surveillance endoscopy recurrence will objectively be defined by the Higaki criteria for recurrence: tumor appearing within a clear resection scar; tumors with convergent folds; and tumors nearby a clear resection scar (within 5 mm).(39) Targeted biopsies will be taken for histological confirmation; in case of an apparently healed normal scar without evidence of recurrence, biopsies will be taken from the basis and 3 from the edges of the scar to detect occult recurrent neoplasia.

The following standardized discharge criteria will be applied in all participating hospitals: normal intake of nutrition; normal mobility; absence of fever (<38°C); and stable hemoglobin level during 1 day (<1 mmol/L) in case of rectal blood loss.

Additional outcome measures:

1. Complications: subdivided into procedural (during treatment) and delayed complications (after ending the procedure); and further subdivided into major (requiring additional surgery) and minor (requiring endoscopic or medical intervention) complications.

   During admission patients will be monitored for complications. In case of same day discharge from the hospital patients will be called by telephone 1 day after the procedure whether adverse events have happened. Two weeks after the intervention, a research nurse will contact the patient by telephone again and ask for occurred complications.

2. Generic and disease-specific health related quality of life will be measured at baseline, 2 weeks, 3 months, 6 months, 1 and year follow-up by the SF-36 and Wexner score (for incontinence) questionnaires.

3. Measurement of anorectal functional outcome by anal manometry and rectal volumetry (barostat) before and 3 months after treatment.

4. Costs of TEM and ESD from a societal perspective, based on primary data (see economic evaluation section).

5. Patient preferences regarding TEM or ESD will be measured at the end of follow-up by a structured questionnaire to enable a discrete choice experiment addressing the burden of care, burden of complications, prognostic uncertainties, and recurrence rates of both treatments.

Sample size calculation Assuming a baseline recurrence rate of 6% for both TEM and ESD (average recurrence based on a systematic review) and considering an upper limit of 10% for ESD to be non-inferior, with a β-error of 0.2 and α-error of 0.05, 60 patients are needed per randomization group.

Since ESD is known to be effective in even more than 2 attempts, an upper limit of 10% seems reasonable, whereas higher recurrence rates would lead to many additional procedures which renders this strategy impracticable and probably not cost-effective.

Patient recruitment Consecutive eligible patients will be recruited at the outpatient clinic in the participating centers by the involved physician (surgeon/gastroenterologist). All patients fulfilling the above-mentioned criteria will be informed about the study by the physician. After consent is given, central randomization will take place web-based and patients will be treated according to the study protocol. Patients unable or refusing to provide informed consent will be treated according to current clinical guidelines.

Data-analysis Since the main outcome of this study is the neoplasia recurrence rate, i.e. proportion of patients with recurrent disease, the Chi-square test will be used to compare the intervention groups (TEM versus ESD). Since the event of recurrence, and not time to recurrence, is the most important indicator for treatment failure, Kaplan Meier methods will not be used. The complication and mortality rates will be compared in the same manner. The number of days not spent in hospital as additional primary outcome measure will be compared by the Wilcoxon rank sum test.

Differences between the intervention groups in continuous outcome measures (e.g. Wexner incontinence scale, quality of life questionnaires, and manometry/volumetry) will be tested by the student's t-test or Wilcoxon rank sum test, where appropriate. All analyses will be carried out primarily on an intention-to-treat basis.

Time schedule Patients will be recruited for this study from October 2009 until December 2010 in all participating centers. All randomized patients will be followed up for 12 months (until December 2011). The last 3 months (until March 2012) will be used for statistical analysis and reporting of data.

Inclusion Criteria:

1. Diagnosed with a large non-pedunculated rectal adenoma (sessile or flat) with a largest diameter of ≥2 cm (estimated by an opened resection snare of 20 or 30 mm).
2. The lower and upper borders of the adenoma are located at ≥2 cm and ≤15 cm from the anal verge, respectively.
3. Biopsies of the lesion did not show malignant neoplastic tissue on histopathological evaluation; only lesions with low or high grade dysplasia are suitable for inclusion.
4. During flexible video endoscopy there are no signs of endoscopic suspicion for submucosal invasive cancer (Kudo pit pattern type V; excavated/depressed type morphology; fold convergence; or large smooth nodule >1 cm in a flat lesion) (33). In case of doubt, patients will undergo EUS as described at (2).
5. In case doubt remains after flexible video endoscopy, endoscopic ultrasonography (EUS) of the rectal adenoma should exclude invasion into the submucosal layer and exclude pathological lymphadenopathy (lymph nodes >1 cm). When pathological lymph nodes are present, fine needle aspiration will be performed to exclude lymph node metastasis (N+ disease).
6. If not performed already, total colonoscopy will be done to detect and remove all synchronous colonic adenomas or cancers first. Cecal intubation must be confirmed by identification of the appendiceal orifice and ileocecal valve.
7. The general health condition of the patient permits general anesthesia (ASA- classification I-III).
8. Absence of non-correctable coagulopathy (international normalized ratio >1,5, or platelet count <90 × 109/l).
9. Patient age of 18 years or older.

Exclusion Criteria:

1. Preoperative histologically detected malignancy
2. Previous anorectal surgery
3. Contraindications to general anaesthesia