**1. Introduction**

86 Updated Topics in Minimally Invasive Abdominal Surgery

[3] O'Rourke N, Fielding G: Laparoscopic right hepatectomy: surgical technique. J

[4] Dagher I, Di Giuro G, Lainas P, Franco D: Laparoscopic right hepatectomy with selective

[5] Han HS, Cho JY, Yoon YS: Techniques for performing laparoscopic liver resection in various hepatic locations. J Hepatobiliary Pancreat Surg2009; 16: 427-432. [6] Gayet B, Cavaliere D, Vibert E, Perniceni T, Levard H, Denet C, Christidis C, Blain A, Mal F: Totally laparoscopic right hepatectomy. Am J Surg2007; 194: 685-689. [7] Takasaki K, Kobayashi S, Tanaka S, Saito A, Yamamoto M, Hanyu F: Highly

[8] Takasaki K: Glissonean pedicle transection method for hepatic resection: a new concept of liver segmentation. J Hepatobiliary Pancreat Surg1998; 5: 286-291. [9] Cho A, Yamamoto H, Nagata M, Takiguchi N, Shimada H, Kainuma O, Souda H, Gunji

[10] Cho A, Asano T, Yamamoto H, Nagata M, Takiguchi N, Kainuma O, Souda H, Gunji H,

[11] Cho A, Yamamoto H, Kainuma O, Souda H, Ikeda A, Takiguchi N, Nagata M: Safe and

[12] Machado MA, Makdissi FF, Galvão FH, Machado MC: Intrahepatic Glissonian

[13] Topal B, Aerts R, Penninckx F: Laparoscopic intrahepatic Glissonian approach for right hepatectomy is safe, simple, and reproducible. Surg Endosc 2007; 21: 2111.

anatomically systematized hepatic resection with Glissonean sheath code

H, Miyazaki A, Ikeda A, Matsumoto I: Safe and feasible inflow occlusion in

Miyazaki A, Nojima H, Ikeda A, Matsumoto I, Ryu M, Makino H, Okazumi S: Laparoscopy-assisted hepatic lobectomy using hilar Glissonean pedicle transection.

feasible extrahepatic Glissonean access in laparoscopic anatomical liver resection.

approach for laparoscopic right segmental liver resections. Am J Surg 2008; 196: 38-

vascular exclusion. J Gastrointest Surg2009; 13: 148-149.

transection at the hepatic hilus. Int Surg1990; 75: 73-77.

laparoscopic liver resection. Surg Endosc2009; 23: 906-908.

Gastrointest Surg2004; 8: 213-216.

Surg Endosc2007; 21: 1466-1468.

Surg Endosc2011; 25: 1333-1336.

42.

Recent improvements in cross sectional imaging, chemotherapy and advances in the techniques of liver resection have resulted in rates of 5 year survival approaching 60% for patients with colorectal liver metastasis. Historically liver resection was perceived as a formidable operation but now liver resection is safe and specialist centres should expect low mortality rates in the region of 1-2%1,2. Consequently, many more patients are now referred for liver resection and its indications are continually being revised and expanded.

At the same time there have been many advances in minimally invasive laparoscopic surgical techniques so much so that laparoscopic liver resection (LLR) is becoming an increasingly popular option amongst laparoscopic enthusiasts. Indeed the first laparoscopic liver resection was described nearly 20 years ago for focal nodular hyperplasia3. In a recent review by Nguyen and colleagues 4,5 over 3,000 laparoscopic liver resections have now been reported in various series and meta-analyses 6 <sup>7</sup> 8. Despite this enthusiasm doubts still remain over its more widespread application because of the risks of complications and whether there is any patient benefit 9-11. The latter is still very difficult to demonstrate in the absence of any well designed randomized controlled trials. Like laparoscopic cholecystectomy that came before, it is now very unlikely that any well designed Randomised controlled trials (RCT) will ever be performed. Perhaps the most important RCT that should have been done is outcome after laparoscopic left lateral resection versus open resection. Yet for laparoscopic enthusiasts the advantages are so obvious they would now be very reluctant to offer open resection in a trial setting. The situation is very different for major resections e.g. right hepatectomy where any advantage is still very difficult to demonstrate. In this situation a RCT would be difficult to design as few centres regularly perform this operation and large numbers would be needed because of high rates of conversion and recruiting patients with tumours distributed in such away that they can be resected laparoscopically.

## **2. Indications and contra-indications**

#### **2.1 International consensus - The Louisville statement**

In 2008 a consensus meeting was convened in Louisville to discuss the position of LLR amongst some of the worlds leading hepatobiliary surgeons. This was a very important development and the following guidelines were suggested as follows11:

1. LLR can be performed safely in specialized centres with results comparable to those achieved after open resection

Laparoscopic Liver Surgery 89

Fig. 1. Hepatic Adenoma ideally placed for Laparoscopic resection

A. Those patients with metastasis confined to liver

B. Those patients with concomitant extra-hepatic disease

i. Unilobar or bilobar disease ii. Single or multiple metastases

resectable local recurrence

adrenal).

should be considered resectable and falls into one of the following groups;

remnant function or equivalent to at least two liver segments

i. CRLM in the presence of resectable or ablatable pulmonary disease

Surgery for LLR should be divided in two broad categories, a) Those patients with metastasis confined to liver and b) those patients with concomitant extra-hepatic disease. Essentially all patients with CRLM who have had radical treatment for their primary CRC

iii. Remnant liver is approximately 20-30%. Total liver volume (TLV) dependent on

ii. CRLM in the presence of resectable isolated extra-hepatic disease e.g. spleen, adrenal or

iii. CRLM in the presence of resectable invasion of adjacent structures (e.g. diaphragm,

With respect to extra-hepatic disease. Elias et al. have reported overall 5-year crude survival rates of 28% when hepatic and extra-hepatic disease are both resected in a curative manner, however in this situation it must be accepted that an R0 resection will not be possible in 50% of patients 24. More importantly, the presence of extra hepatic disease does not appear to influence outcome when resection is complete along with the liver metastases 25 . Nevertheless it cannot be denied that there are few long term survivors in the presence of peritoneal disease 26. Certainly these types of patients should be carefully evaluated by open


Other areas of discussion focused on patient safety and contraindications with the following guidelines being suggested


#### **2.2 Benign liver tumours**

Paired comparisons between laparoscopic and open resection for benign tumours have not been frequently reported 16 <sup>17</sup> 18. A few series are dedicated to LLR for benign tumours only but these can be subdivided into solid or cystic 19 20. Most studies report outcomes in series mixed for both benign and malignant tumours 21. The largest series of LLR for benign tumours have been reported by Koffron et al. (n=177) 22. Forty seven were hepatic adenomas the others being made up of haemangiomas (n=37), FNH (n=23) and liver cysts (n=70). It is not clear in this article what the indications for resection were. Most centres report predominantly resection of malignant tumours. From the Newcastle series of 69 patients; 28% constitute benign lesions and 72% malignant. The most common benign lesions include hepatic adenoma **(Figure 1),** symptomatic FNH (or where there was diagnostic doubt), biliary cysts, angiomyolipoma, haemangioma, biliary haematoma and polycystic liver disease. In our experience most of these lesions were resected in patients with a known diagnosis of colorectal carcinoma where there was diagnostic doubt regarding a liver lesion despite cross sectional imaging by CT and MRI and in some cases contrast enhanced ultrasound. It is important not to expand the indications for resection just because it can be done laparoscopically. In general for benign tumours most report less morbidity (including incisional hernias), shorter hospital stay and faster time to oral intake 19.

#### **2.3 Malignant liver tumours**

Although there have been many reports of LLR for malignant tumours being resected including hilar cholangiocarcinoma 23 and neuroendocrine/carcinoid tumours, for the purposes of this chapter discussion will concentrate on the most commonly resected malignant tumours e.g. CRLM and HCC.

#### **2.4 Colorectal**

One of the disadvantages of LLR for CRLM is that all patients have had previous surgery and initial dissection can be tedious because of adhesions. Especially when patients have had a previous right hemicolectomy or cholecystectomy. Indeed in one patient in the authors series LLR was abandoned after 3 hours of dissection and failure to progress.

2. The main indications are for both symptomatic benign and malignant tumours the latter being predominantly Hepatocellular carcinoma (HCC) and liver metastasis

3. It is important that the indications for resection of benign liver tumours are not

4. Harvested grafts for living donation should only be performed in very specialised

Other areas of discussion focused on patient safety and contraindications with the following

Paired comparisons between laparoscopic and open resection for benign tumours have not been frequently reported 16 <sup>17</sup> 18. A few series are dedicated to LLR for benign tumours only but these can be subdivided into solid or cystic 19 20. Most studies report outcomes in series mixed for both benign and malignant tumours 21. The largest series of LLR for benign tumours have been reported by Koffron et al. (n=177) 22. Forty seven were hepatic adenomas the others being made up of haemangiomas (n=37), FNH (n=23) and liver cysts (n=70). It is not clear in this article what the indications for resection were. Most centres report predominantly resection of malignant tumours. From the Newcastle series of 69 patients; 28% constitute benign lesions and 72% malignant. The most common benign lesions include hepatic adenoma **(Figure 1),** symptomatic FNH (or where there was diagnostic doubt), biliary cysts, angiomyolipoma, haemangioma, biliary haematoma and polycystic liver disease. In our experience most of these lesions were resected in patients with a known diagnosis of colorectal carcinoma where there was diagnostic doubt regarding a liver lesion despite cross sectional imaging by CT and MRI and in some cases contrast enhanced ultrasound. It is important not to expand the indications for resection just because it can be done laparoscopically. In general for benign tumours most report less morbidity (including

Although there have been many reports of LLR for malignant tumours being resected including hilar cholangiocarcinoma 23 and neuroendocrine/carcinoid tumours, for the purposes of this chapter discussion will concentrate on the most commonly resected

One of the disadvantages of LLR for CRLM is that all patients have had previous surgery and initial dissection can be tedious because of adhesions. Especially when patients have had a previous right hemicolectomy or cholecystectomy. Indeed in one patient in the authors series LLR was abandoned after 3 hours of dissection and failure to progress.

expanded (e.g. asymptomatic tumours where there is no diagnostic doubt)

1. The contraindications for LLR should be the same as those for open resection.

3. The presence of dense adhesions and failing to progress after prolonged dissection

(colorectal-CRLM) and in determinant liver lesions.

guidelines being suggested

**2.2 Benign liver tumours** 

**2.3 Malignant liver tumours** 

**2.4 Colorectal** 

malignant tumours e.g. CRLM and HCC.

2. Other contraindications include;

5. Tumour too large to manipulate

4. Tumour adjacent to a major vascular structure

6. The need for a portal lymphadenectomy.

centres and should be scrutinized in a world registry 12 13-15.

incisional hernias), shorter hospital stay and faster time to oral intake 19.

Fig. 1. Hepatic Adenoma ideally placed for Laparoscopic resection

Surgery for LLR should be divided in two broad categories, a) Those patients with metastasis confined to liver and b) those patients with concomitant extra-hepatic disease. Essentially all patients with CRLM who have had radical treatment for their primary CRC should be considered resectable and falls into one of the following groups;

A. Those patients with metastasis confined to liver


With respect to extra-hepatic disease. Elias et al. have reported overall 5-year crude survival rates of 28% when hepatic and extra-hepatic disease are both resected in a curative manner, however in this situation it must be accepted that an R0 resection will not be possible in 50% of patients 24. More importantly, the presence of extra hepatic disease does not appear to influence outcome when resection is complete along with the liver metastases 25 . Nevertheless it cannot be denied that there are few long term survivors in the presence of peritoneal disease 26. Certainly these types of patients should be carefully evaluated by open

Laparoscopic Liver Surgery 91

specialist who removes the primary in one sitting. Up to 25% of patients may present in this way 32. Nonetheless there are no significant publications with any reasonable numbers to draw on any useful conclusions as to whether there is any benefit with combined laparoscopic procedures 33 34. Minimally invasive techniques have obvious advantages over two major laparotomies in a short space of time. With advances in chemotherapy more patients are now becoming operable with their primary still in situ as there liver disease can be controlled. This cohort is becoming increasingly more common and challenging 35 36. Generally these patients have either laparoscopic right hemicoloectomy or laparoscopic anterior resection with excision of either a solitary or unilobar metastasis. A further group includes those patients who have major colonic resection with clearing of a single lobe and then further downstaging chemotherapy prior to definitive resection by a second open liver resection. Recent reports have suggested no significant differences in post-operative morbidity or mortality or 5 year survival rates in those patients with synchronous disease who need a minor hepatectomy with colonic resection 37 38 . In patients who require a major hepatectomy, a test of time, to enable an assessment of the biological behaviour of the disease and to provide adjuvant treatment, is still sensible. Although simultaneous laparoscopic major liver resection e.g. right hepatectomy along with major colonic resection e.g. anterior resection have been successfully described 39 <sup>40</sup> the authors would not recommend this without a careful assessment of the patients fitness

because of the need for prolonged anaesthesia beyond 5 hours.

Fig. 3. Colorectal metastasis with the primary colonic tumour still in situ ideal for

simultaneous laparoscopic

surgery and not by LLR. The easiest patients to consider for LLR are those with disease confined to a single segment who ideally have a solitary metastasis in the anterior segments (IVb, V and in some cases VI) (**Figure 2.)** or in the left lateral segment (group Ai or Aii). Laparoscopic posterior sectionectomy has been described but they are significantly more challenging 27-29.

Fig. 2. Colorectal metastasis in segment VI for laparoscopic resection

Patients with extra-hepatic disease (group B) fall into a very difficult group as resection of extra-heptic disease may require more advanced laparoscopic skills which could be more easily dealt with by open surgery. The temptation to laparoscopically resect a single lesion and then perhaps laparoscopically ablate a more difficult lesion should be avoided and open surgery performed.

With respect to nodal disease, regional metastasis to peri-hepatic lymph nodes deposits should not be regarded as a contraindication to open resection but does reduce long-term outcome. Recent studies suggest up to 20% of patients will have hepatic nodal involvement at the time of resection 30. It is very difficult to evaluate this laparoscopically. Resection of nodes involving second tier nodes (i.e.celiac nodes) is far more controversial and offers no survival benefit. Another problem highlighted by the MSKCC group is the ability to identify which lymph nodes are involved during open surgery. Routine sampling of lymph node stations and lymphadenectomy is unnecessary and time consuming, without any evidence of benefit. The best approach is selective sampling based on intra-operative assessment and pre-operative imaging 31. Again performing this laparoscopically would not be advisable.

One of the main advantages of LLR, in our experience, has been its use with synchronous tumours **(Figure 3).** LLR can be performed at the same time with a laparoscopic colorectal

surgery and not by LLR. The easiest patients to consider for LLR are those with disease confined to a single segment who ideally have a solitary metastasis in the anterior segments (IVb, V and in some cases VI) (**Figure 2.)** or in the left lateral segment (group Ai or Aii). Laparoscopic posterior sectionectomy has been described but they are significantly more

Fig. 2. Colorectal metastasis in segment VI for laparoscopic resection

Patients with extra-hepatic disease (group B) fall into a very difficult group as resection of extra-heptic disease may require more advanced laparoscopic skills which could be more easily dealt with by open surgery. The temptation to laparoscopically resect a single lesion and then perhaps laparoscopically ablate a more difficult lesion should be avoided and open

With respect to nodal disease, regional metastasis to peri-hepatic lymph nodes deposits should not be regarded as a contraindication to open resection but does reduce long-term outcome. Recent studies suggest up to 20% of patients will have hepatic nodal involvement at the time of resection 30. It is very difficult to evaluate this laparoscopically. Resection of nodes involving second tier nodes (i.e.celiac nodes) is far more controversial and offers no survival benefit. Another problem highlighted by the MSKCC group is the ability to identify which lymph nodes are involved during open surgery. Routine sampling of lymph node stations and lymphadenectomy is unnecessary and time consuming, without any evidence of benefit. The best approach is selective sampling based on intra-operative assessment and pre-operative imaging 31. Again performing this laparoscopically would not be advisable. One of the main advantages of LLR, in our experience, has been its use with synchronous tumours **(Figure 3).** LLR can be performed at the same time with a laparoscopic colorectal

challenging 27-29.

surgery performed.

specialist who removes the primary in one sitting. Up to 25% of patients may present in this way 32. Nonetheless there are no significant publications with any reasonable numbers to draw on any useful conclusions as to whether there is any benefit with combined laparoscopic procedures 33 34. Minimally invasive techniques have obvious advantages over two major laparotomies in a short space of time. With advances in chemotherapy more patients are now becoming operable with their primary still in situ as there liver disease can be controlled. This cohort is becoming increasingly more common and challenging 35 36. Generally these patients have either laparoscopic right hemicoloectomy or laparoscopic anterior resection with excision of either a solitary or unilobar metastasis. A further group includes those patients who have major colonic resection with clearing of a single lobe and then further downstaging chemotherapy prior to definitive resection by a second open liver resection. Recent reports have suggested no significant differences in post-operative morbidity or mortality or 5 year survival rates in those patients with synchronous disease who need a minor hepatectomy with colonic resection 37 38 . In patients who require a major hepatectomy, a test of time, to enable an assessment of the biological behaviour of the disease and to provide adjuvant treatment, is still sensible. Although simultaneous laparoscopic major liver resection e.g. right hepatectomy along with major colonic resection e.g. anterior resection have been successfully described 39 <sup>40</sup> the authors would not recommend this without a careful assessment of the patients fitness

because of the need for prolonged anaesthesia beyond 5 hours.

Fig. 3. Colorectal metastasis with the primary colonic tumour still in situ ideal for simultaneous laparoscopic

Laparoscopic Liver Surgery 93

This modality is the work-horse of all imaging techniques in the pre-operative planning phase for LLR. Present generation triple phase multi-detector CT scanning technology enables image acquisition during a single-breath hold, of the entire chest and abdomen and pelvis. The improved resolution results in excellent detection of lesions in solid organs and enables better local, regional and distant staging. The other advantage of CT scanning is the high incidence of detection of lesions in the lung, liver and pelvis, when intravenous contrast is used with arterial or venous phase scanning. Slice thickness or maximum collimation should be 3- 5mm. The sensitivity for detecting a metastatic lesion approaches 80%, which increases to 90% when CT angiography is used, however lesions less than 1 cm in size are liable to be missed 56. Contrast enhanced helical CT is the investigation of choice in the initial evaluation of liver tumours assessing response to chemotherapeutic agents and

Magnetic resonance imaging has an extremely high sensitivity in identifying and characterizing small lesions within the liver. In addition patients are not exposed to radiation but the procedure is far more expensive and labour intensive. One of its limitations is the identification of extra hepatic disease. The technique is very sensitive to respiratory artefact and this can limit its resolution in certain patients who are unable to hold their breath for a sufficient length of time. Contrast agents such as gadolinium and the liver specific super magnetic iron oxide result in very high sensitivities in diagnosing small (less than 1 cm) liver metastases 57 and differentiating between potentially malignant and benign liver lesions (e.g. FNH, adenoma etc). Usually MR imaging is utilised just prior to resection, in order to identify small lesions not visualised by conventional CT scanning but

Intra-operative ultrasound (IOUS) is an essential pre-requisite for assessment of the liver prior to commencement of liver resection. IOUS allows for mapping of the major vascular and ductal structures in relation to the metastasis and aids in planning the final approach to resection. It also serves as a guide in confirming the accuracy of the plane of dissection. However following chemotherapy, when fatty change supervenes and in the presence of cirrhosis, identification of small iso-echoic masses becomes poor, decreasing the sensitivity of IOUS. IOUS must be used before, during and at the end of resection in order to keep R1 resection rates as low as possible. It is also important to leave an adequate margin around the tumour and to mark the margins prior to commencing parenchymal transaction. This is also useful to avoid coning as it is very difficult to estimate the depth of a tumour without

One of the overlooked contra-indications for LLR is the patients inability to withstand a prolonged pneumoperitoneum especially with major resections e.g. right hepatectomy. Results of left lateral liver resection suggest that resection time can be comparable to open.

**3. Imaging** 

**3.1 Computed tomography** 

for post-operative surveillance for tumour recurrence.

**3.2 Magnetic resonance imaging** 

this is not universally routine.

**3.3 Intra-operative ultrasound** 

measuring the dimensions.

**4. Anaesthesia** 

Two-thirds of patients undergoing liver resection for CRLM will develop recurrence of their disease within 2 years 32. One third will manifest with liver only disease and a small proportion of them will be suitable for repeat liver resection 41. Technically repeat liver resections are demanding. However long-term survival is similar to those following initial liver resections for open resections 42 43 . In a series of 60 third hepatectomies 43 complication rates were similar to those having first and second hepatectomies with no obvious survival disadvantage. Five year survival rates of 32% have been reported after open resection. Multivariate analysis suggests a curative resection (R0) as the most important predictor of improved survival after open resection. There are no studies reporting repeat LLR but these are likely to be technically more challenging. Further studies are needed to evaluate repeat LLR in terms of survival rates and complications.

#### **2.5 Hepatocellular carcinoma**

The treatment of HCC covers a broad spectrum including surgical (Resection, Ablation or Liver Transplant-LT) and non-surgical treatments (Sphere therapy, TACE, Sorafenib). Mortality after liver resection in large series of non cirrhotics are now around 3%. Yet in large volume centres in the east, mortality after resection for HCC in cirrhotics is now approaching zero. Substantial refinements in the surgical techniques have played an important role including the development of liver "hardware" such as ultrasonic dissectors, low CVP anaesthesia, hepatoduodenal compression (Pringle's manoeuvre) and vascular staplers have all contributed to reducing blood loss, post-op morbidity and mortality 44.

There is no doubt that the results of LT for primary HCC have improved dramatically in the last decade following the publication of the Milan criteria by Mazzaferro et al. in 1996 45. Consequently more patients with HCC are being referred for consideration of LT and the management of these patients on the ever expanding waiting list present an interesting cohort of patients to discuss. With this in mind bridging treatments such as resection, chemo-embolisation or ablation by RFA are becoming increasingly important The clinical characteristics after such treatments are also important in terms of predicting overall prognosis.

One of the disadvantages of resection is tumour recurrence as some suggest that this can hinder subsequent LT 46 yet this has not been substantiated by others and in terms of technical difficulty is no different to re-transplantation for other indications 47. To avoid this problem there is a niche for the development of LLR which can reduce morbidity and have an impact on curative intent as a potential bridging treatment. Resection can be useful as a bridging treatment if patients are Childs A with a low MELD score, have a small tumour <3cm without any obvious macroscopic tumour thrombus 44. Overall 3 year survival rates in patients with Child's A cirrhosis can be as high as 93% 48 for segmental resections. Segmental resections are best performed given the risks of recurrence with non segmental resections due to microscopic satellite nodules that are not easily visualised by intraoperative ultrasound. Comparisons of LLR with open resection for HCC in cirrhotic patients are favourable 49 <sup>50</sup> <sup>51</sup> 52 but the main advantage of LLR is a shorter hospital stay and less blood loss. LLR is also less likely to lead to problematic adhesions if LT is required at a later date. Numerous single centre 49 <sup>50</sup> <sup>53</sup> 54 and multi-centre series 55 have published their series of LLR in patients with HCC and cirrhosis confirming it is safe and reproducible without oncological compromise or survival.
