Despite the advantages of the laparoscopic approach for colorectal surgery, this approach has some limitations such as loss of the 3D vision, limitations in the freedom degrees of the surgical instruments, the amplification of the physiological tremor and the “fulcrum” effect.

The implementation of robotic technology avoids this disadvantages and improves the ergonomics of the surgeon[26]. Uhrich et al[27] proved that the uncomfortable positions during laparoscopic surgery increase surgeons fatigue and iatrogenic injuries.

The development of robotic surgery started in the mid-80s of the 20^th century mainly focused in the development of tele-surgery. The FDA approved the use of Da Vinci system (Intuitive Surgical, Sunnyvale, CA, United States) in 2000, and nowadays is the only robotic system available for minimally invasive surgery.

The first article published about robotic colonic surgery using Da Vinci Surgical System was from Weber et al[28] reporting a right colectomy and a sigmoidectomy for benign diseases. At the same time, other authors published their first robotic colonic surgery case reports[29-31].

In 2004, D'Annibale et al[32] reported 53 patients that had undertaken robotic colonic resections. In this group, 22 were resections for oncological reasons. They concluded that the operative and postoperative results were similar to those obtained with conventional laparoscopic instruments. Braumann et al[33] published the first 5 robotic colonic resections performed in Germany in 2005.

Pigazzi et al[34] and Hellan et al[35] published the first article reporting robotic rectal resections in 2006. These authors reported the first series of 39 consecutive resections for rectal cancer, concluding that this technique is safe and feasible.

In Asia, the first robotic rectal surgery with total mesorectal excision was performed by Baik in June 2006[36]. The same author published the first extended resection with hysterectomy[37]. Ng et al[38] reported the first robotic abdominoperineal resection in Hong Kong.

Few data are available regarding the value of robotic colonic resections but the results seem to be similar to those reported by conventional laparoscopic approaches[39]. No final conclusions can be made at this moment.

There are not many articles regarding robotic rectal resections in the scientific literature. Hopefully, the ROLARR (Robotic vs Laparoscopic Resection for Rectal cancer) trial results will be available in a few months and will give some valuable information.

Baek et al[40] analyzed the results of 64 rectal resections for cancer with no operative mortality, a mean blood loss of 200 mL, a conversion rate of 9.4% and a leak rate of 7.7%. The mean of harvested lymph nodes was 14.5, the distal margin was 3.4 cm and the circumferential margin was negative in all cases. They found a local recurrence in 6 patients after an interval of 23 mo. The disease free survival rate was 73.7%.

Luca et al[41], performed a study analyzing sexual and urinary function in 74 patients after robotic rectal resections concluding that robotic instrumentation helped to preserve sexual and urinary function after total mesorectal excision.

Biffi et al[42], studied the blood loss during robotic rectal resections and reported that only one case of their series of 49 patients required blood transfusion, by contrast with patients with open surgery who required blood transfusion in 12 cases of 105.

Shiomi et al[43], in a study with 113 consecutive robotic rectal resections, 12 of them having T4 tumors, observed no conversion and no hospital mortality. The frequency of Clavien III/IV complications was 2.7%. They concluded that robotic instrumentation was helpful in performing advanced dissections with a very low morbidity and conversion rates.

deSouza et al[44] compared the results obtained with robotic surgery in 36 consecutive cases against those obtained in 46 cases with laparoscopic assisted surgery using a hand port for the splenic flexure. The authors conclude that robotic total mesorectal excision is feasible and safe, and comparable with open surgery in terms of perioperative and anatomopathological results. There was a statistically significant difference in the tumor location, with more mid and low rectal tumor in the robotic group.

Kwak et al[45] in another retrospective study concluded that the results of robotic surgery were comparable to those obtained with a laparoscopic group. This study was performed by a single surgeon with a huge experience in minimally invasive surgery for the treatment of rectal cancer. He recognizes a selection bias in this study and concludes that prospective, multicentric and randomized studies are necessary.

A very interesting study from Kang et al[46], compared three groups of patients with mid and low rectal tumors treated with either open, laparoscopic or robotic approach. They observed that the robotic group had a faster postoperative recovery with a lower hospital stay, less pain and better specimen quality. The disease free survival rate was similar in all groups three years after surgery.

Fernandez et al[47], retrospectively compared a group of patients treated with a robotic approach vs a group treated with a laparoscopic approach. They performed a low anterior resection in all cases with low anastomosis. They observed no difference in blood loss, postoperative morbidity or surgical specimen quality but, nevertheless, they recognized that the robotic group had lower tumors with a more advanced disease and more chemo radiation. The conversion rate was 17% in the laparoscopic group and 8% in the robotic group.

Patriti et al[48] performed a study comparing 29 patients with robotic rectal resections against 37 treated with a laparoscopic approach with a one-year follow-up. They obtained similar results in both groups with a higher conversion rate in the laparoscopic group: 7% vs 0%.

Lin et al[49], performed a meta-analysis concluding that robotic surgery is clearly superior in terms of conversion rate. Another meta-analysis of studies comparing robotic vs laparoscopic surgery for rectal cancer performed by Trastulli et al[50] suggested that the robotic surgery group had a statistically significant difference in conversion rate without significant differences in operation time, hospital stay, postoperative morbimortality or surgical specimen quality. The meta-analysis performed by Ortiz-Oshiro et al[51] had similar conclusions.

In the systematic review of the literature performed by Scarpinata and Aly[52] excluding the studies that referred to colonic resections, they suggested that there is evidence that robotic surgery may offer a better short term results, mainly in obese or male patients. It also may be better in those cases with previous radiotherapy and lower tumors. There was no evidence of any difference in terms of leakages, circumferential margins or preservation of the autonomous function.

A study performed by a Taiwanese group[53], compared the postoperative results of 64 patients with ultralow anterior resection and intersphincteric dissection. Twenty-eight patients had undertaken a conventional laparoscopic approach and 36 a robotic approach. They found statistically significant differences in terms of surgical time - longer in the case of robotic surgery - and in the number of definitive stomas, which was 46.4% in the laparoscopic group vs 19.4% in the robotic surgery group. The authors conclude that this kind of procedure is feasible and safe with robotic instrumentation, with better functional outcomes and that surgical time will diminish as the experience of the surgeons increases.

Casillas et al[54], analyzed the results of robotic colorectal surgery performed in a single institution by a single surgeon. They compared 200 laparoscopic cases vs 144 robotic cases. They observed a shorter hospital stay and a lower complication rate in the case of robotic surgery.

Park et al[55], performed a prospective study with 217 patients that undertook minimally invasive surgery for rectal cancer. 133 patients had robotic surgery while 84 had a conventional laparoscopic approach. There were statistically significant differences in favor of the robotic approach in terms of hospital stay and conversion rates (0% vs 7.1%). Overall survival rate and disease free survival rate were similar in both groups with a 5-year follow-up. Saklani et al[56] have reported similar results in a 3-year follow up study.

In a recently published meta-analysis, Xiong et al[57] made a comparative analysis between laparoscopic and robotic rectal surgery in terms of safety and efficacy. They identified 8 studies with an overall number of 1229 patients, 554 robotic cases vs 675 laparoscopic cases. The authors concluded that the robotic approach is safe and feasible, but they did not find statistically significant differences in circumferential margins or in sexual function after surgery.

In summary, rectal robotic excision is feasible and safe, is comparable to laparoscopic surgery in terms of short and long-term outcomes, with some advantages such as shorter hospital stay, lower conversion rates and better functional results. Some particular conditions such as lower rectal tumors, male and/or obese patients or locally advanced rectal tumors may be indications that could benefit the most from the robotic approach.

The use of robotic instrumentation in rectal surgery requires not only training in surgical technique but also training in the use of the robotic system. This training has the specific handicap of the loss of the tactile feedback. The procedure is performed from a console far from the patient. This requires an excellent coordination between the surgeon and the surgical assistant.

The advantages that robotic technology provides make, for an expert surgeon in open surgery, less necessary the previous training in conventional laparoscopic surgery. The surgeon expert in conventional laparoscopic procedures has to make a specific training in robotic surgery. No differences have been observed in the learning curve for robotic surgery in surgeons with previous training in laparoscopic surgery vs those without that training.

Giulianotti et al[58] observed that the robotic surgery learning curve was short for easy procedures as suture knotting or instrument use. They also observed that the training for advanced surgical procedures required previous experience in open and laparoscopic surgery.

Very few studies analyze the learning curve in robotic surgery. Bokhari et al[59] estimated following the Cumulative Summation (CUSUM) technique that 50 robotic rectal procedures are necessary for a surgeon to be proficient in this procedure. Sng et al[60] have recently published an article reporting a multiphasic learning curve. In the first phase (around 35 initial cases) the surgeon performs selected easy cases. In the second phase, with a worse CUSUM, the surgeon performs more complex cases (to 100 cases). Finally the surgeon enters in a consolidation phase.

Buchs et al[61] have reported that the learning curve can be reduced using simulation in an animal/cadaver model or through the visualization of video clips or attending to courses.

In a recent study, Byrn et al[62] compared their initial 43 robotic rectal surgery cases with the following 42 cases observing a significant reduction of surgical time and costs.

Other considerations concerning training are made at the end of the paper.

One of the most criticized aspects of robotic colorectal surgery is the increase of cost per procedure. Cost analysis may vary depending on the criteria we use and the items we analyze. Rectal cancer process is very long and the analysis of costs of this process may vary a lot if we analyze just the perioperative period or a 5-year period including stoma cost, quality of life or local recurrence. The cost analysis also varies a lot depending on the health care system characteristics. There are not yet high quality articles that are conclusive about the cost issue.

Delaney et al[63] reported a significant increase of in-hospital costs with robotic colorectal surgery: 2946 dollars per laparoscopic procedure vs 3721.5 dollars per robotic resection. This increase was mainly attributed to the increase in intraoperative costs.

However Rawlings et al[64] did not find any statistically significant difference when they analyzed intraoperative, staff and surgical time costs.

Park et al[55] compared robotic with laparoscopic rectal cases with 150 laparoscopic rectal cases and concluded that the robotic approach increased the perioperative costs. These authors recommended cost effectiveness studies including long-term results, oncological results and functional outcomes. Bodner et al[65] have reported similar results.

Some limitations of the MIE can be overcome by the aid of the robotic systems, which provide some advantages such as a tridimensional view of a field selected for the surgeon instead of the assistant-, the 7 degrees of freedom allowing movements similar to open surgery, the tremor suppression resulting in a better precision, and a better ergonomics which leads to less surgeon fatigue. All of these advantages are even greater in small surgical fields, with few instrument exchanges as the thoracic phase of an esophagectomy is.

The first robot-assisted MIE using the Da Vinci system was published by Kernstine et al[93] in 2004. A few short series, between 6 and 47 cases, have been reported since then, always with cervical anastomosis[94-99]. The first robot-assisted Ivor Lewis series were published from 2013 to now, reporting 22[100], 17[101], and 50[102] cases, respectively, all with the patient in lateral decubitus. In 2014 our group published the first series of robot-assisted Ivor Lewis in prone decubitus with intrathoracic manual anastomosis[103]. We feel that the prone position makes the dissection and lymphadenectomy easier, in an optimal field. We had no respiratory complication in 39 cases, although a stapled anastomosis, either transthoracic or transoral, is more difficult. The robotic assistance makes a manual intrathoracic anastomosis easier and allowed us to use the prone position and its potential advantages without flawing short and long-term oncological outcomes, as we recently reported in a series of 21 cases[104].

Despite these potential advantages, the evidence to show any possible superiority of the robot-assisted MIE over either open esophagectomy or conventional thoracoscopic esophagectomy is still very limited. Since 2012 a single center controlled randomized trial has been ongoing in the Netherlands to compare the robot-assisted and open esophagectomy[105], with a recruitment prevision of 112 patients - 56 per arm and a follow-up of 5 years. However, for stronger evidence, multicentric trials with a large number of patients is needed.

For tumors in early stage and distal location, literature is profuse in retrospective studies, case series and comparative studies but there are seldom randomized controlled trials (RCTs) comparing Laparoscopic distal gastrectomy (LDG) with open gastrectomy and they have a limited case number[115-122]. Despite the limitations, LDG appears to have advantages over open gastrectomy in terms of postoperative pain, recovery of gastrointestinal and pulmonary function, hospital stay and return to normal activity. The complexity of a proper lymphadenectomy - especially if a D2 is mandatory, and the concerns about oncological safety has slowed down its generalization[123].

Additional evidence has been provided by several meta-analysis[124-127], some of them with more than 3000 patients including advanced stages[128-131] and the most recent with 2144 cases[132]. All coincide in the LDG perioperative advantages - blood losses, overall morbidity, hospital stay - and that oncological results are not inferior to those of open surgery, although more operative time is spent.

Currently two multicentric phase III RCTs are ongoing, one from the Japan Clinical Oncology Group[133] and another from the Korean Laparoendoscopic Gastrointestinal Surgery[134]. The joined recruitment prevision is more than 2300 patients and the results are expected for 2015.

At the moment, and after the international expert group meeting, the LDG is accepted for distal tumors cT1-2 cN0 and the laparoscopic total gastrectomy (LTG) for proximal tumors cT1 cN0, although no consensus was achieved for other stages[135].

Literature data concerning LDG in locally advanced tumors (stage II and III) are even scantier. Only two RCTs have compared MIG and open gastrectomy[136,137]. Two meta-analysis, have been published: the first included 7 studies with 174 laparoscopic and 278 open gastrectomies[138], and the second analyzed 10 studies with 495 laparoscopic and 544 open gastrectomies, both with D2 lymphadenectomy[139]. Both coincide in that the minimally invasive approach in advanced gastric cancer is associated with a longer operative time, but less blood losses, pain, postoperative complications and hospital stay. Similar lymph node number and overall survival were found in the two approaches.

A recently published phase II prospective clinical trial with 157 patients concluded that laparoscopic gastrectomy with D2 lymphadenectomy for advanced gastric cancer is technically feasible and safe, with acceptable morbidity and mortality rates[140].

More high quality RCTs comparing open and laparoscopic gastrectomy are needed, as well as multicentric studies. Even so, some aspects will probably not able to be applied to the West due to the above mentioned clinical and histological differences.

The Laparoscopic total gastrectomy (LTG) spreading has been slower worldwide, even in Korea and Japan, due to the need of an esophagogastric anastomosis, which is technically demanding, and because of the low incidence of the proximal gastric cancer in the East. The first relatively large series appeared in 2009, such as that of Shinohara et al[141] with 57 patients or the multicentric study from South Korea by Jeong et al[142] of 131 laparoscopic-assisted total gastrectomies. Both conclude that LTG is feasible and safe - morbidity of 31% and 19%, respectively with no mortality, it is possible to retrieve sufficient lymph nodes - 46 and 35, respectively - although with long operative times - 4.5 h on average.

In the West, the European pioneer groups in this field, such as Dulucq et al[143] or Huscher et al[144] published a series of only 8 and 11 LTG, respectively. In 2013, a specialized center, the MSKCC published a series of 17 cases[145].

Three recent meta-analysis[146-148] also suggest that, in skilled hands, LTG has better perioperative outcomes than the open procedure in terms of blood losses, pain, resumption of oral intake, hospital stay, with no inferiority concerning lymph node retrieved and survival. However, the operative times are longer.

A phase II multicentric prospective trial (KLASS-03) in patients with stage I gastric cancer is currently ongoing, with the aim of assessing perioperative morbidity and mortality of LTG by comparison with the open procedure[149].

The main advantages of the robotic systems have already been discussed. In the case of gastric cancer, these potential advantages lie in a better precision for the lymphadenectomy and an improved skill for intracorporeal anastomosis[150]. Hashizume et al[151] published the first Robot-assisted gastrectomy (RAG) in 2002, but because the procedure is technically complex and the equipment is expensive, the spreading has been slow. Several groups have compared the RAG with the laparoscopic and open techniques. The Yonsei University group published in 2009 the initial experience with 100 patients, extended to 236 cases more (73% subtotal and 27% total) in 2011 of RAG compared with 591 laparoscopic (81% subtotal and 19% total). They concluded that RAG seems to have better short-term results with comparable oncological outcomes[152,153]. Other published series until 2012 support these conclusions, although the largest one reports less than 40 cases[154,155]. The first meta-analysis, also from 2012, compared 268 RAG with 650 laparoscopic gastrectomies[156]. Significant differences were not found in overall morbidity, hospital stay, or number of lymph nodes retrieved.

Interesting, to make the most of the benefits of the robotic assistance, the morbidity due to anastomotic leak must be minimal. A study analyzing postoperative complications in 5839 patients (4542 open, 861 laparoscopic and 436 robotic gastrectomies) concluded that, even in expert hands, minimally invasive techniques are associated with an increased risk of anastomotic leak by comparison with open gastrectomy, although the overall morbidity and mortality rates were similar[157].

Since 2012, seven new meta-analysis have been published, analyzing between 404 and 762 RAG patients[158-164]. All concluded that RAG was associated with lower blood loss and shorter hospital stay, with an adequate lymphadenectomy, by comparison with the laparoscopic and open gastrectomy, although with longer operative times.

More high quality studies are needed to clearly define the role of RAG.

Robot-assisted bariatric surgery (RABS) has been used since 1998, when a gastric band was put from the distance[194], and shows some advantages over the laparoscopic techniques, which have some limitations related with a poor ergonomics due to a limited instrument mobility due to the abdominal wall width, and hepatomegaly. RABS suppresses the position port limitation as well as the physiological tremor and confers a better ergonomics[195]. The three dimensional view allows a more precise dissection[196] and a decrease in blood loss. A shorter learning curve by comparison with conventional laparoscopy has been claimed[197].

Several studies report that RABS is safe with lower complications than the laparoscopic techniques. Edelson et al[198] compared 287 robotic and 120 laparoscopic gastric banding cases, and they did not find any significant difference in intraoperative or postoperative complications, hospital stay or operative time; the time was, however, significantly lower in patients with a BMI > 50 kg/m² operated on through a robotic approach (91 min vs 103 min). Fourman et al[199] reported similar findings in a literature review of RABS which included gastric banding.

The gastric bypass has been used since 2000, with satisfactory outcomes[200]. Lower complication rates than those of the laparoscopic approach, without mortality or anastomotic leaks have been published[199,201-205]. Others also report significantly less anastomotic failures with RABS[206]. Skilled teams have achieved similar operative times and even shorter[199,206,207], with comparable long-term results concerning weight loss and comorbidity improvement.

In a review of 1686 patients comparing laparoscopic and robotic bypass, similar results were found regarding anastomotic leaks, postoperative complications, operative time and hospital stay[208]. However, an advantage was found in a decrease of the anastomosis stenosis rate at 6 mo. Most groups coincide in that anastomosis leak is lower with RABS, although without reaching statistical significance.

The vertical gastrectomy or sleeve gastrectomy (SG) has become in one of the most popular bariatric procedures due to its effectiveness in reducing weight. Overweight losses as high as 61% after 24 mo have been reported[209], as well as comorbidity improvement such as diabetes resolution in 47%-66%[210,211]. Other advantage are the lower operative time needed, the shorter hospital stay and the lower frequency of complications, by comparison with the laparoscopic GB. Since the use of robotic surgery has been limited to the most complex procedures - revisional surgery and gastric bypass, there are only a few studies on SG[199,211-217] which do not show any significant difference concerning complication frequency - stenosis, bleeding-, mortality or hospital stay. One study reports a lower fistula frequency and a shorter stay in the robotic cases, but without statistical significance[218]. This technique, when performed with robot assistance, is associated with longer operative times and is more expensive, and thus is controversial its generalized use. However, it is proposed as a way to learn robotic skills before performing the gastric bypass.

The Scopinaro biliopancreatic diversion (BPD) and the biliopancreatic bypass with duodenal switch (DS) are more effective than the gastric bypass in achieving weight loss and improvement of obesity associated comorbidities[219,220] but due to their complexity, higher complication rates and the need of nutritional control, are the least used. The first results of these procedures performed through a laparoscopic approach were published by Ren et al[221] in 2000, with a 2.5% mortality. In a systematic review, a 30-d mortality of 0.1% for restrictive procedures, 0.5% for gastric bypass and 1.1% for BPB/DS were reported 59. Sudan et al[222] reported a series of 59 robotic BPB/DS without mortality, anastomotic leaks, bleeding, sepsis or pulmonary thromboembolism.

The main criticism to RABS is its high cost, as well as the longer operative time, especially because of the preoperative docking time needed. This however, can be minimized with increasing experience of the surgical team.

Although in general formal pancreatic resection is recommended for most tumors, enucleation can be performed for neuroendocrine neoplasms if tumor size does not exceed 2 cm and if no findings of malignancy are detected on preoperative staging[264]. Pancreatic tumor enucleation can be easily performed by laparoscopy with excellent morbidity and mortality outcomes[265].

This is the most frequently performed type of laparoscopic pancreatectomy for both benign and malignant diseases[266,267].

Distal pancreatectomy can be accomplished with or without splenic preservation. Splenectomy could adversely influence oncologic long-term outcomes, in addition to predispose to infectious complications. As a result, efforts to preserve the spleen should be done in case of benign or low grade malignancies, provided that splenic vessels are not involved with the tumor. More controversial is splenic preservation in case of adenocarcinoma. Whatever the case, if the spleen is to be preserved, two techniques are used[264,268]: in the first, section of the splenic vessels both at the level of transection of the pancreas and at the splenic hilum is performed, leaving the short gastric vessels as the only blood flow supply. In the second technique, the splenic vessels are preserved by meticulous ligation of all the branches reaching the pancreas. However, patency of splenic vein - although not that of the artery - can be compromised after a laparoscopic DP in as high as 35%[269]. The frequency of splenic infarction and appearance of gastric varices is higher in case of splenic vessel section[268].

Blood losses have been reported to be significantly lower in case of a laparoscopic approach by comparison with open procedures[270-273]. Localizing small tumors by laparoscopy or laparoscopic ultrasonography can be difficult and this leads to conversion in many cases. Other causes of conversion can be bleeding or difficult structure identification which lead to conversion in 17%-30% of cases[270-272]. Obesity is significantly associated to conversion[271].

The complication rate has been reported significantly lower in laparoscopic cases[271,274], although the severity grade was similar than in open procedures[263]. Other papers showed similar rates[272,273]. Tran Cao et al[267] studied a nationwide database and compared the short-term outcomes of 382 minimally invasive distal pancreatectomy (mainly laparoscopic and some robotic) with those of open distal pancreatectomy. They found a significant reduction in overall perioperative morbidity among patients undergoing minimally invasive surgery, including a significant decrease in hemorrhagic complications and postoperative infections in laparoscopically treated patients. Results of five meta-analyses[266] support these findings concerning overall perioperative morbidity, although clinically relevant pancreatic fistula frequency was significantly lower in laparoscopic cases only in one of the analyzed studies.

Patients converted from laparoscopic to open surgery have significantly more severe complications than those with not converted[271] which reflects the need of proper selection for a laparoscopic approach. The reported mortality ranges 0%-1%[267,271,273].

Some have reported operation times significantly longer in laparoscopic than in open procedures[271], but others find similar duration[273].

The hospital stay has also been reported to be significantly lower with a laparoscopic approach[27,270,271,273].

A recently published meta-analyses[275], comprising data of 3701 patients, all of them from non-randomized studies, confirmed most of the above mentioned findings: superiority of laparoscopic over open DP in terms of blood loss, time to first oral intake, and hospital stay. Mean operation time, mortality -0.4% in DP-morbidity and safety showed no difference. However, data concerning oncologic radicality and effectiveness are limited.

Most articles report a mixture of benign and malignant diseases as indication of distal pancreatectomy (DP). As a result, to reach conclusions concerning safety and oncological outcomes when dealing with adenocarcinomas is difficult. A few papers compare open and laparoscopic approaches only in case of pancreatic adenocarcinoma. Kooby et al[270] performed a multicenter matched analysis of 23 laparoscopic procedures compared with 189 open procedures and found no significant differences in positive margin rates, number of nodes examined or overall survival. The median follow-up was only 10 mo and, thus, it is premature to conclude that the long-term results are as safe as in open procedures. Magge et al[272] compared 28 patients with 34 operated on by an open approach. They found not significant differences in margin-negative resection (open, 88%; laparoscopic, 86%) and median lymph node clearance. Also, no significant differences were found for overall survival or risk of local recurrence.

A recent paper compared laparoscopic an open DP performed for neuroendocrine tumors[274]. The laparoscopic procedure showed no mortality, a significant lower complication rate and shorter hospital stay. No significant difference was found concerning margin involvement, long-term survival, and overall costs.

In summary, DP is feasible, safe and reproducible for most laparoscopy skilled surgeons. Concerning its use in case of adenocarcinoma, it also appears as safe as open DP but more studies analyzing long-term results are needed.

Theoretical advantages and disadvantages of robotic surgery have been discussed earlier. The published papers dealing with robotic DP are still scanty.

Zureikat et al[276] report 83 robotic DP with no mortality and a low rate of severe complications, although with 43% cases of pancreatic fistula. The average operative time was 256 min and 2% required conversion.

Waters et al[277] compared 17 robotic against 32 open and 28 laparoscopic DP. Cystic tumors predominated among the indications of robotic resection. In this group, longer operative times, similar blood losses, no mortality, shorter hospital stay and lower hospital costs were found. Higher rates of splenic preservation were achieved in comparison with the open and laparoscopic approaches. Daouadi et al[278] compared 30 robotic DP and 94 laparoscopic DP patients. The robotic DP group included more adenocarcinomas. Significant differences in blood loss, hospital stay, or morbidity were not found. However, there was a statistically significant 36% increase in R0 resection with robotic surgery, as well as more lymph nodes harvested. Also the conversion rate was lower.

The study by Kang et al[279] included 20 robotic and 25 laparoscopic DP for benign and borderline malignant lesions. Although the intent was to preserve the spleen, the overall rate of splenic preservation was very low. A higher splenic preservation rate was achieved with the robotic approach. This took longer operative times and had increased costs compared to laparoscopic cases.

In summary, although feasible and safe, robotic DP has not yet shown clear advantages over laparoscopic DP.

This technique can be applied for the treatment of benign or borderline lesions of the pancreas situated in the pancreas neck. Although uncommonly performed, this technique is feasible by laparoscopy[265-280].

A review from 2013[281] collected 51 published cases operated on through total laparoscopy or robotic assistance. Pancreatic reconstruction was done with a Roux-en-Y pancreato-jejunostomy, or pancreato-gastrostomy. The procedure was long with a mean time of 356 min. Blood loss was minimal in most cases. Mortality was nil, but morbidity was high, mainly due to pancreatic fistula (46%).

When performed by a robotic approach, the procedure is associated with long operative times (mean: 394 min) and a 23% of severe complications, although without mortality[276].

Three techniques are currently employed for LPD: pure laparoscopy (PL), hand-assisted (HA) laparoscopy, and laparoscopic-assisted (LA) surgery. In contrast with DP, where only resection is to be performed, the pancreaticoduodenectomy (PD) requires a complex resection as well as pancreatico-jejunal, hepato-jejunal and gastro-jejunal anastomoses. As a result, although feasible, the difficulty of the procedure makes that experience with laparoscopic pancreaticoduodenectomy is limited to a few case series.

Gumbs et al[282] analyzed most reported cases until 2011 comprising 285 patients, although only 32% had pancreatic adenocarcinoma. The most important findings for the entire cohort were a 9% conversion rate, 2% mortality, 48% morbidity and a length of stay of 12 d. Margin involvement was found in 0.4% with an average of retrieved lymph nodes of 15. The mean disadvantage was a long operation time ranging from 263 to 750 min.

Direct comparison between laparoscopic and open PD has been performed by some groups. The first[273], which included 60% of patients with malignant tumors - pancreatic, biliary, ampullary - found significantly lower blood losses, and hospital stay as well as more lymph node harvested in the laparoscopic group. However, the average operative time was significantly increased. There were no differences in overall complications, pancreatic fistula, delayed gastric emptying, and resection margin involvement. The main finding of the second study[283] was a significant reduction in blood losses and, thus, the transfusion need in the laparoscopic patients although it did not show relevant differences in morbidity.

Asbun and Stauffer[284] compared 53 laparoscopic and 215 open PD. They found significant differences favoring laparoscopic PD blood losses, transfusions, length of hospital stay, length of ICU stay (P < 0.001), and number of lymph nodes retrieved although the operative time was significantly longer. The rates of overall complications, pancreas fistula, delayed gastric emptying, margin involvement were not significantly different.

Kuroki et al[283] compared 20 laparoscopic and 31 open PD. They also report a significant reduction of blood losses and transfusion need but they did not find significant differences in morbidity.

Song et al[285] report a comparison between 137 laparoscopic and 2055 open pylorus-preserving PD for periampullary tumors. A shorter hospital stay and a lower analgesic consumption was seen in the laparoscopic cases. No other difference was found including complications, blood losses margin involvement or lymph nodes retrieved.

In a recent systematic review of several series[286], the operative time averaged 464.3 min (338-710 min) with conversion in 9.1% of patients. The mean estimated blood loss was 575 mL. Average mortality was 1.9% and morbidity ranged between 18.1% and 64.2%, with pancreatic fistula ranging between 4.5% and 52.3% of patients. An average of 14.4 lymph nodes were retrieved and 4.4% of cases had marginal involvement.

In summary, laparoscopic PD is feasible and as safe as the open procedure in highly skilled hands and high-volume centers. Long operative times are needed. Decreased blood losses seem to be the main advantage. Concerning its use in case of adenocarcinoma, more studies are needed.

Published experiences concerning robotic PD are even less common. Zureikat et al[276] report 132 cases with a 90 d mortality rate of 3.8% and a frequency of severe complications (III-IV) of 21%. The frequency of pancreatic fistula was 17%, although only 3.7% were of grade C. Conversion was needed in 8%. The procedure took long operative times (mean 527 min), and although decreased with improved experience, the last cases of the series lasted over 400 min.

In the systematic review by Boggi et al[286], the robotic PD used up more operative time and had higher pancreatic fistula rate by comparison with pure laparoscopic PD (but not laparoscopically assisted PD). However, significant differences were not found concerning overall morbidity or mortality (2.7%, 1.1%, and 2.4% for pure laparoscopic, laparoscopically assisted, and robotic PD, respectively).

Another systematic review analyzed several non-randomized studies comparing open and robotic pancreatectomies, most of them PD[287]. It included 137 cases of robotic PD and 203 open PD. The median conversion rate was 10%. Overall complications, reoperation rate, and margin involvement were significantly lower in robotic group, with no significant difference in postoperative pancreatic fistula, and mortality. In one study included in this review[288] involving only PD, a significant increase in the R0 resection rate, with 100 % of patients in the robotic group compared with 87% in the open group.

A recently published paper[289] comparing robotic and open PD, report similar findings such as lower blood losses, shorter hospital stay, longer operative times with no significant difference in morbidity, mortality, as well as R0 resection rate and number of lymph nodes retrieved in case of malignancy. Patients with pancreatic adenocarcinoma of both groups had similar overall and disease-free survival.

In summary, the same as laparoscopic PD, robotic PD is feasible in highly skilled hands and high-volume centers. Although the very short published experience makes to reach firm conclusions difficult, it suggests that robotic PD is as safe as laparoscopic PD with some advantages over open PD. However, the former is more time-consuming.

With the arrival of laparoscopic surgery to the surgical departments in the 90s, the surgical resident training has suffered a huge transformation with a decrease in its autonomy. This has resulted in the need of additional training to obtain the confidence and maturity of judgment. As a result, the number of fellowships has dramatically increased during the last ten years. This change has enlarged the surgical training[290].

Laparoscopic surgery training programs have evolved in several ways in the different countries. Sweden or The Netherlands have national programs limited to basic procedures (cholecystectomy) using training techniques as virtual reality[292]. In the United States and Canada, the surgical training method is structured and the program of the American College of Surgeons (ACS) demands the inclusion of basic training in the Fundamentals of Laparoscopic Surgery (FLS) in the residency programs[293]. Spain has a limited training of two 3-d courses for the residents, one basic and another for advanced training[294]. In Latin America, the project of Laparoscopia Avanzada Práctica (LAP), pretends to bring near a big number of countries to basic programs in laparoscopic training[295].

One of the most important aspects of the residency program is that training has to be based in a curricular model. Based on the premises that experts indicate[296], these curricula have to be: (1) Endorsed by an accredited training institution (ACS) with a clear message regarding how this surgical training will be; (2) If simulation is used as a training tool, this demands a new and different approach of the instructors; (3) Training has to be integrated with clinical practice; (4) Adequate simulators have to be used in the correct timing and they should be validated for training; (5) The features and different types of surgical simulators are continuously changing. This has to be taken into account in order to plan the update of these simulators; (6) Residents must have reserved time to use simulators in their training; and (7) Financing has to be sustainable using business principles.

Virtual reality (VR) simulators provide surgical training without supervision in a safe environment for both patient and trainees. Skills obtained in the VR simulator can be transferred to the theatre. However, evidence is only limited to basic surgical skills and laparoscopic cholecystectomy. There is no evidence yet of the effect of VR simulators on advanced laparoscopic procedures[297]. The introduction of haptic feedback in these VR simulators has not increased the validation of laparoscopic surgical competences[298]. In summary, all the trails that compare the training using VR simulators and standard laparoscopic training in the theatre observe a higher performance after training with VR, confirming that current training using structured simulation is more efficient than traditional training[291]. Supported by these findings, countries such as Sweden or The Netherlands have established a structured training for laparoscopic cholecystectomy based in the use of VR simulators. This has allowed these countries validate resident competency before being trained in the theatre[299,300].

Animal models for resident training in advanced laparoscopic techniques (Nissen, Colectomy) are the more realistic models, but they are limited in some countries because of religious beliefs or laws. All these reasons are behind the substitution of these animals by synthetic models or even by simulated models with “ex vivo” viscera[301]. Without any doubt, the animal model most frequently used is the porcine model, mainly in colorectal surgery training[302]. Despite this, it is not possible to define nowadays when competency is reached with an animal model, if this model is the best one for laparoscopic training and if what is learned in this model is translated into clinical practice.

There are few studies about the use of cadaveric human models[303]. These models have been used for training in laparoscopic skills, such as cholecystectomy, and have demonstrated the increase of capacitation by comparison with VR simulators. We believe that this model has to be reserved for advanced laparoscopic procedures (colorectal or bariatric surgery) that require a huge degree of realism. Leblanc et al[304] assessed the use of fresh human cadavers for surgical training in laparoscopic sigmoidectomy. They observed that the use of this model improved clinical practice in terms of dissection, traction/counter traction, eye-hand coordination, suture, bleeding control and theatre time comparing with the use of a VR simulator. Palter et al[305] studied the sequencing of VR simulators with human cadaver models for training in colorectal laparoscopic surgery. He added a cognitive module in order to help the participants to understand the procedure and how to plan and execute a right colectomy or a sigmoidectomy. He observed that this training approach improved the technical knowledge and the performance in the theatre in comparison with the traditional training during the residency program.

Hybrid models are those that use a complex robotized mannequin together with the abdomen/thorax of a live animal. These models allow us a high laparoscopic realism simulation while we adjust the cardiologic/respiratory parameters of the robotized mannequin. This way we are able to simulate for example a coronary event during the simulated laparoscopic procedure.

In other opportunities[306], scenarios can be created to train the laparoscopic surgery team in a simulated theatre with a hybrid patient (SimMan 3G; Laerdal) and a laparoscopic VR simulator (Lap Mentor Express, Simbionix). These authors observed that the global assessment of the team showed a high qualification. Powers et al[307] observed that these simulation models let us discriminate between the technical and non-technical abilities in residents and experienced surgeons. The target of this innovative multidisciplinary simulation is to identify the problem and to start with the adequate solution by the surgical team.

Zendejas et al[308] observed in a recent review that whenever simulator is used, it has huge advantages in laparoscopic or open surgery training comparing with no simulator use. These authors also observed that adding the use of simulators to the traditional training is more effective than the use of traditional training alone.

Following the experience of authors such as Haluck et al[296], we think that the University Hospital Marqués de Valdecilla resident training meets a number of requirements[309]: (1) Our curriculum in laparoscopic procedures is developed during the full residency period. We think like Sadideen et al[310] that the first steps in surgical skill training have to be done outside the theatre and that practice is the most important thing to achieve automaticity in surgical skills. In the clinical environment, the needs of the patient prevail over the needs of the trainee; (2) The curriculum is compound by 19 laparoscopic modules; this allows the trainee to progressively gain competence as he learns. At the same time, each year modules are more technically advanced; (3) We support training in a simulated environment and advance in the learning curve during simulation. We have observed that doing gastrointestinal anastomosis and colonic resections this curve is shorter and increases patient safety[311]; (4) Working with the most realistic models in each training phase, we observed that an important part of the initial training can be performed with low cost animal “ex vivo” models. Some examples are gastrointestinal anastomosis, cholecystectomy or gastric bypass; (5) The modular curriculum covers not only technical skills competencies but also teamwork competencies during crisis in laparoscopic surgery procedures. Teamwork competencies are trained in hybrid simulators; (6) The training process of the resident is assessed with Global Rating Scales of Operative Performance. Even though we also apply this assessment to the clinical practice, it is very difficult to move this assessment to the professional competence[312]; and (7) The ACS accredits this training program and the center where it is developed.

We think that progress is very difficult and it may be necessary a trial/error system that let us advance. Learning from other programs that have tried, failed or succeeded may be a key point[313].

For most of the groups around the world, the most important challenge is defining how simulation based training can be implemented and improve the training system. There are many questions to be answered, as we cannot say with scientific evidence in which degree simulated training improves results or quality in clinical practice and patient safety.

Simulators have to improve their benefits and design, and objective measures have to be developed so we can say in which degree the trainee acquires the adequate clinical competencies that can be translated in to the theatre[314].

Mattar et al[315] observed that general surgery residency inadequately prepares trainees for fellowship results of a survey of fellowship program directors.

Fellowship programs are well established in countries such as the United States, Canada, United Kingdom or Australia. On the other hand, they have not had a good degree of implementation in the rest of Europe or in Eastern countries. These fellowship programs are advanced laparoscopic surgery training programs that have their own problems. Fellowships coexist with residency programs in the same institution and the continuous advance of new surgical techniques and the complexity of the surgical equipment make that some of the training is short in time or objectives. These reasons are making fellowships more and more specific as the recent fellowships in robotic colorectal surgery or rectal surgery is[316].

The development of national training programs in advanced laparoscopic surgery, such as the one developed in the United Kingdom[317] in laparoscopic colorectal surgery (Lapco) have shown good results in terms of short- and long-term training and have had a positive impact on the trainee learning curves[318,319].

There are an increasing number of short length training courses in advanced laparoscopic surgery. They are usually limited to 3-5 d and they include live procedures performed by expert surgeons in most of the cases. Some theoretical knowledge is also given either during the course or on-line and, in some cases, the trainees have the opportunity to practice the technique on an animal or cadaver model. These courses have been mostly developed in colorectal, bariatric and upper gastro-intestinal surgery.

The impact of these courses on the training of the participant will depend on his previous experience in laparoscopic or open surgery. This way, we see that surgeons that come to a course with previous laparoscopic experience posteriorly implement the acquired knowledge during the course in a 60%-70%. Without this experience the degree of implementation is under 25%[320,321]. Kinoshita et al[322] have demonstrated in Japan that after a training course in gastric laparoscopic surgery the number of laparoscopic gastrectomies increased in a 50% in the participating institutions. Participants answered to the survey saying that they felt improvement in their surgical skills in 100% of the initial procedures. On the other hand, Brunckhorst et al[323] say that there is very poor evidence concerning the training value of live operations and that very few high quality studies have been performed in this field.

Our point of view is that these 3-5 d courses provide knowledge and skills that help the trainee in starting laparoscopic surgery in an already established unit. These courses also help surgeons in sharing experience with experts.

In summary, we can say that current literature consistently proves the positive impact of simulation in theatre time and the scores in predefined performance but, however, this is not enough to ensure the transfer of these lab acquired skills to the theatre.

We are living a huge technological advance in all social scopes and also in surgical training[324]. In this context, the acquisition of knowledge is progressively moving to e-learning platforms that reduce classroom hours. This system includes interactive feedback with the instructor, assessment of the procedure and interaction with other participants. This system may also be combined with VR training[325].

VR simulators for laparoscopic surgery have been importantly improved with haptic technologies. In the near future, it may even be possible to import real 3D images to VR software. This may allow the trainees to perform real operations in virtual surgical fields that look like the real ones. Modelling this imported images, the trainee may even be able to work in fields with anatomical variations[326].

“Virtual cadavers” based on 3D images reconstructed from computerized tomographies will replace human cadaver and animal models. It will be possible to create huge libraries with this “virtual cadavers”. The exposure to multiple scenarios during the same basic procedure will make easier the trainee progress from competency to expertise.

Tele-simulation will be possible thanks to this libraries allowing tele-training. Virtual environments as second life (SL) will be used to completely represent a training centre or meeting room. These environments are already available in the market and can make possible the interaction of scenarios, patients, VR simulators, lectures or videoclips. Surgeons will be able to build their own virtual clinic or whatever they may need to simulate according to the level of competency that is trained[327].

Tele-mentoring[328] using robots as RP-7 (RP-7, Intouch Health, Santa Barbara, California) makes possible active mentoring of the trainees with verbal instructions or changing position of the instruments/camera when needed. It also makes possible a passive mentoring just with verbal instructions. This tele-mentoring seems to be a valuable tool for training minimally invasive procedures.

Now is the moment when all this separate tools, laparoscopic surgery, tele-presence, VR, digital image, networking… join together making tele-surgery possible. A surgeon is able to be miles away from the theatre and assist a surgical procedure.

Robotic surgery has been progressively incorporated to advanced laparoscopic procedures and has made those procedures easier. Robotics will facilitate the training of those surgical procedures. Mixing VR or virtual libraries with the surgical console of the robotic surgical systems will make it possible to train the procedure before doing it in the real world. It will make training a particular procedure in a particular patient possible.

In summary, we should imagine a surgeon being trained in his work environment in complex minimally invasive procedures by anther surgeon that is “on-line”. Robotic surgical systems will be present in daily work and training.