The prevention of perioperative complications begins with an accurate assessment of the patient in order to optimize their clinical status and detect liable treatment conditions. Currently, there is no scale of global perioperative risk; hence patient’s clinical risk and surgical risk evaluation are done separately. In general, liver resections are classified as high-risk surgeries due to functional implications, the possibility of bleeding, surgical position, and the duration of the procedure. Liver resections are also associated with a risk of cardiovascular (CV) events of around 5%. The ASA score (scale of the American Society of Anesthesiologists) is an assessment of the clinical status of the patient, not a risk scale, so it must be complemented with an assessment of CV and respiratory risk, especially in older patients and those with a related medical history. Acute renal failure in the postoperative period has been associated with increased mortality at 90 days posthepatectomy [12, 13]. Therefore, it is necessary to evaluate and optimize renal function before surgery, and to modify perioperative management, controlling renal perfusion and avoiding nephrotoxic drugs. Diabetic patients show an increased risk of infection, thrombosis, CV, and renal failure [14, 15]. Furthermore, liver regeneration may be reduced, with the consequent increased risk of PHLF, especially in patients who undergo major liver resections.

Advanced age is associated with an increased prevalence of comorbidities and disabilities. The aging process involves a decrease in the physiological reserves of vital organs (heart, liver, and kidney), therefore increasing the operative risk. In recent years, the concept of fragility has been developed, and is defined as the biological syndrome where the reserve and resistance to stressors reduce functionality progressively [16]. This deterioration leads to an increased vulnerability for adverse outcomes. The elderly patient group is not a homogeneous population as chronological age does not always reflect the biological age. Therefore, the global evaluation of these patients should be carried out including not only comorbid conditions, but also geriatric syndromes and frailty (weight loss, weakness, low resistance, etc.). The fragility syndrome has been associated in several studies with an increased frequency of postoperative complications, prolonged hospitalization, and death [16]. The treatment strategy should therefore be tailored according to the existence of both patient- and liver-related operative risks.

In the treatment of liver tumors the surgeon must follow two contradictory objectives: (1) a complete resection of the tumor(s), and (2) the preservation of as much parenchyma as possible to prevent the development of PHLF, which is the most severe complication after major hepatic resections. Moreover, postoperative mortality has been related to PHLF in approximately half of patients [17]. Risk factors include previous parenchymal damage, long-course chemotherapy regimens, size and functional status of the liver remnant, and intraoperative ischemia developed by the surgeon. For this reason, it is essential to evaluate the quality and quantity of the remaining liver prior to surgery, as well as to detect portal hypertension at an early stage.

Several advances in imaging methods have led to a more accurate preoperative staging at both local and systemic levels, and thus a better planning of treatment strategies. Imaging modalities in patients with CLM should provide information on the number, size, and location of metastasis, its relationship with vascular and biliary structures, and the presence of lymph node metastases in either the hepatic hilum or other locations. The presence of extrahepatic disease should also be evaluated. Currently, the most used methods for staging are magnetic resonance imaging (MRI), computed tomographic scan (CT-scan) and positron emission tomography scan (PET-scan) [24]. These last two studies can be merged for a better anatomical resolution. Despite having a lower sensitivity than MRI, the CT-scan is the most used method for the study of the liver and extrahepatic metastasis, remaining as the first-line diagnostic tool in most centers [24]. Modern multi-slice equipment with high definition as well as different phases of intravenous contrast makes possible reconstructions in any anatomical plane, and even three-dimensional vascular reconstructions (Angio-CT). An MRI with diffusion-weighted images has a higher sensitivity for the detection of CLM, reaching 100% in lesions larger than 1 cm [24]. It also makes it possible to distinguish the different soft tissues, so it is often used to define the metastatic nature of unspecific tumors observed in CT-scan. The systematization of the study protocol is very important to achieve excellent results. This includes axial planes, T2 sequences, T2 fat suppression, in- and out-phases, diffusion, and dynamic study with intravenous contrast. While MRI is used in most of the centers as second diagnostic line, it should be used as first line especially in patients who have received chemotherapy (in which TC has low sensitivity) or to detect small or missing lesions. The PET-scan has high sensitivity and specificity comparable to or even greater than CT-scan for initial staging of patients with liver metastasis. This study is useful to detect extrahepatic disease and to stage patients with recurrence in whom a new resection is planned. However, it has a weak anatomical definition and should be complemented by specific imaging studies.

Another key piece of information to define the surgical strategy provided by modern preoperative imaging concerns the liver anatomy itself, and the relationship between the metastatic disease and the different structures of the portal pedicle or hepatic veins. For example, the presence of an accessory hepatic vein (e.g., segment 6) could change the surgical approach, allowing certain parenchymal preserving strategies [25]. Therefore, the knowledge of normal liver anatomy and its variations is essential to plan the strategy, based on the information provided by the imaging studies.

The application of preoperative systemic chemotherapy in patients with CLM may be used in to ways: (1) as neoadjuvant therapy for resectable tumors, or (2) as conversion therapy to transform initially not resectable disease into resectable disease. The opportunity to convert a patient with an initially unresectable disease to resectable has further expanded the pool of patients who may benefit from surgery. This type of treatment requires intense collaboration between the team of surgeons and oncologists, who should reassess the treatment’s response every 2–3 months and define the right time for liver resection. Liver metastasis should be resected as soon as they become technically resectable. This makes it possible to stay under the therapeutic window for response in order to avoid progression during treatment, which is associated with poor prognosis, but also to shorten the treatment duration and prevent liver damage.

Even though patients with response or stable disease under chemotherapy are those who benefit the most from surgical resection, recent evidence suggests that patients under progression but with less than three lesions, none of them larger than 5 cm and with a CEA of less than 200 ng/ml can have similar results to those without progressive disease [26]. Different chemotherapy regimens, associated or not with monoclonal antibodies, have been studied with promising results, reaching high rates of post-treatment R0 resections in selected patients. However, intensive use of chemotherapy is not harmless for the liver parenchyma. In fact, the use of pre-hepatectomy chemotherapy has been associated with higher morbidity and mortality after resection, especially when more than five cycles are administered [27]. The different types of parenchymal injuries possible are associated with the use of specific drugs. The use of 5-fluorouracil has been associated with the development of liver steatosis, while irinotecan with steatosis and chemotherapy-associated steatohepatitis (CASH), which has been found to be a risk factor for postoperative morbidity and mortality [28]. On the other hand, oxaliplatin is associated with sinusoidal obstruction syndrome (SOS), which is related to increased risk of postoperative bleeding [29]. Some studies suggest that combining with bevacizumab (anti-VGF) may protect from this syndrome caused by oxaliplatin [30]. Although the use of anti-EGFR antibodies (cetuximab, panitumumab) has not been associated with increased risk of morbidity, a higher incidence of bleeding, impaired tissue healing and possible alterations in liver regeneration are described for bevacizumab [28]. However, these changes could be minimized if the medication is suspended at least 5 weeks before surgery [31].

Another problem with chemotherapy, more worrisome when using it as neoadjuvant treatment in patients with resectable disease, is the disappearance of metastasis in preoperative imaging. Some authors propose differentiating between those metastases that fade (“vanishing”) in imaging but are visible during laparotomy, and those that are absent in imaging and cannot be detected during surgery (“missing”). Imaging studies have limitations in detecting small and superficial lesions; so the general consensus defines that patients should be explored surgically even if metastases are not visible on preoperative imaging [32]. Every lesion detected during laparotomy through visualization, palpation, and/or intraoperative ultrasonography (“vanishing metastasis”), should be ideally resected or at least treated with a local ablative method. One of the biggest challenges for a liver surgeon is the scenario of a disappearing metastasis on preoperative imaging that cannot be detected during surgery (“missing metastasis”). This scenario should be avoided when using preoperative chemotherapy, since complete tumor necrosis in these missing sites is exceptional (4–20%) and therefore there is a high risk of leaving microscopic disease behind [33, 34]. Given that complete pathologic response has been associated with an increased long-term survival, until a few years ago most experts recommended resection of parenchyma where metastases were located previously [33]. Even though nowadays there is not an absolute consensus, if a metastasis cannot be detected after careful and thorough palpation and intraoperative ultrasonography, blind resection should be avoided, and patients should be carefully followed by CEA and imaging methods in order to detect early recurrence if any.

The evolution of liver surgery and chemotherapy regimens in the last 10 years has reduced the need of large liver resections [35]. Nowadays, a greater number of parenchymal sparing strategies are being performed (Fig. 5.4), which are considered by many the first-choice strategy because it preserves non-tumoral parenchyma, allows repeated resection in case of recurrence, and does not compromise oncological outcomes [25, 35, 36]. Parenchymal sparing resections might be particularly beneficial for patients with a high operative risk for major resection, who would otherwise not be candidates for resection. Intraoperative ultrasound confirms and extends previous findings, becoming essential for intraoperative decision- making [25].

The paradigm shift from large to parenchymal sparing resections was possible mainly due to a modification of the oncological concept of safe resection margins. In the 1990s, a lesion was considered non-resectable if it could not be resected with a margin of at least 1 cm [6]. Later on, not reaching this margin was not a contraindication but a strong recommendation [37]. Conversely, in the last 15 years, various authors have shown comparable results with narrower margins and even with positive microscopic margins (R1) [38, 39]. However, the current consensus is that the thickness of the margin does not modify survival as long as it is negative (R0 resection), although R1 resection alone should not be a contraindication for surgery. Nowadays, unresectable extrahepatic metastases or unresectable primary tumor, prohibitive anesthesiological risk, and medical contraindications to hepatectomy still constitute contraindications for resection.

Despite the fact that parenchymal sparing minor resections are preferable in most patients, in certain cases the treatment of extensive tumor burden and/or unfavourable disease location requires the use of major liver resections, which carries a significant risk of PHLF when the FLR is regarded as insufficient. The old paradigm of resectability began to change with the introduction of different techniques to increase FLR volume, allowing safer major curative resections. Makucchi and coworkers described in 1990 the use of transileocolic portal vein embolization (PVE) to reduce PHLF in patients with hiliar cholangiocarcinoma [40]. The next breakthrough came a decade later, when the group from the Paul Brousse Hospital in Paris introduced a sequential surgical technique known as “two-stage hepatectomy”, which removes multiple liver tumors, allowing the liver to regenerate between the two procedures [41]. Soon after, Belghiti and colleagues in France modified this approach by applying portal ligation and concomitant wedge resection of all tumors on the left side during the first surgery, followed a few weeks later by an extended right hepatectomy [42]. Finally, Jaeck and colleagues developed another two-stage approach for bilateral (predominantly right) disease, consisting of right PVE and resection of tumors located in the left liver during the first stage [43]. In 2009, a group from Seoul reported favourable results with the addition of sequential embolization of the hepatic vein 2 weeks after PVE, inducing a greater contralateral regeneration due to more liver damage than PVE alone [44]. Using this principle, Balzan et al. [45] developed an outflow modulation alternative to induce hypertrophy of insufficient segments 1 and 4. More specifically, in a first surgery a right hepatectomy combined with partial ligation of the left hepatic vein is performed, generating flow redistribution to segments 1 and 4, and therefore inducing its hypertrophy. In a second stage, a left lateral segmentectomy is completed.