It is important to consider the current systemic therapeutic regimen when contemplating elective surgery for an oncology patient. Common cancer therapies and their toxicities are presented in Table 7.2. Chemotherapy has traditionally consisted of cytotoxic agents with broad efficacy across tumour types. The mechanism of action in general relates to inducing DNA damage or inhibiting progression through the cell cycle such that malignant cells die by apoptosis or necrosis [19]. Most chemotherapy agents are given intravenously as single doses separated by 1–4 weeks, to permit recovery from side effects and haematological abnormalities. Combinations of two or more agents are commonly used to improve response rates and circumvent drug resistance. Of note, 5-fluorouracil is often given in continuous infusions lasting days or weeks requiring central venous access. Surgically inserted venous access devices are potentially one of the most common “palliative” surgical interventions. Oral chemotherapeutic agents, such as capecitabine and temozolomide, are given in continuous or interrupted schedules lasting 5 days or more. Capecitabine is an oral 5-fluorouracil derivative which in theory reduces the need for venous access although only if taken as a single agent, which in today’s practice is less common.

The haematological and mucosal toxicities of cytotoxic chemotherapy are likely to be of most concern to the surgeon. Neutropenia and thrombocytopenia may affect the timing of surgery but usually resolve within a 1–2-week time frame. Mucosal toxicities such as chemotherapy-related diarrhoea and cutaneous toxicities are commonly mild but in some cases can be severe and create unfavourable conditions for surgery. It is likely that surgeons today have had experience with the reversible toxicity of standard chemotherapy.

Newer so-called biological or targeted agents have differing toxicity profiles that may also differ in severity and time to recovery compared to what would be considered typical for conventional anticancer therapies. These agents target particular molecular abnormalities present in the tumour that provide a growth advantage. The target for some tumours can be identified pretreatment with mutation testing of pathology specimens, such as the absence of mutations in the KRAS oncogene in colorectal cancer, indicating sensitivity to cetuximab [20]. Immunohistochemical procedures may also identify a target such as human epidermal growth factor receptor 2 (HER2) overexpression in breast cancer, indicating sensitivity to trastuzumab [21]. In other cases the presence of the target is assumed. This is often the case for currently available oral targeted agents which are all inhibitors of the tyrosine kinase domain. This functional domain of various cellular receptors and proteins is mutated or overexpressed in multiple tumour types causing unhindered proliferative signalling and malignant transformation [22]. Agents that inhibit these tyrosine kinase domains have demonstrated efficacy in multiple advanced malignancies (Table 7.3).

Anti-angiogenic therapies that target vascular endothelial growth factor (VEGF) are particularly relevant to the surgeon. They include the anti-VEGF monoclonal antibody bevacizumab and the anti-VEGF tyrosine kinase inhibitors sunitinib, sorafenib and pazopanib. Anti-VEGF agents have been associated with excess bleeding risk, and they may impair wound healing [23–26]. Where possible, surgery should be delayed until a suitable treatment-free period to prevent complications. Data suggest that a 6–8-week interval from the last dose of bevacizumab is sufficient to prevent surgical complications for major abdominal procedures requiring laparotomy [27] and bevacizumab treatment may be commenced 4 weeks after surgery [28]. For minor procedures such as surgical insertion of vascular access devices, a 14-day interval from the last bevacizumab treatment is sufficient [29]. Potent anti-VEGF activity of bevacizumab is still detectable in blood and tissue at 6 weeks from the last treatment, but this is not associated with wound healing complications [30]. There is less evidence regarding the timing of surgery and the use of anti-VEGF tyrosine kinase inhibitors, but the short half-life of these drugs in contrast to bevacizumab suggests that shorter intervals are acceptable. Studies of neoadjuvant administration of these agents in renal cell carcinoma provide evidence that an interval of anywhere between 1 and 16 days from the last dose to major surgery is safe and feasible and that treatment can be restarted 2–6 weeks after surgery if the wound is well healed [31–33].

There is a paucity of data regarding the perioperative safety of agents that target the epidermal growth factor receptor (EGFR) in colon and lung cancer (such as cetuximab, panitumumab, erlotinib) or C-KIT in gastrointestinal stromal tumours (imatinib), but they appear to have less impact on healing and bone marrow function, and thus significant delays in surgery are not considered necessary. Potential cytopenias should, however, be evaluated with a full blood examination and differential count immediately prior to surgery, as imatinib in particular can cause neutropenia [34]. Cetuximab appeared safe when used prior to resection of hepatic metastases from colorectal cancer in a study where surgery was performed a mean of 32 days after chemotherapy (range 12–56 days), with no increase in surgical complications and no evidence of excess hepatotoxicity [35].

Improved outcomes have resulted from the incorporation of multi-agent chemotherapy regimens into the treatment of metastatic colorectal cancer, combining oxaliplatin or irinotecan with 5-fluorouracil or capecitabine [36]. These regimens yield response rates of approximately 40–50 % in patients not previously treated and 10 % in those already treated with chemotherapy, with a median overall survival of around 20 months [1]. The addition of the targeted agents bevacizumab (anti-VEGF) and, for KRAS wild-type tumours, cetuximab or panitumumab (anti-EGFR) has brought about further incremental improvements in response rates and progression-free and overall survival [37]. In general the order of the regimens used is considered less important than exposing a patient to all agents [38], although the potential for hepatic surgery or the presence of an in situ primary lesion does play a role in drug choice and hence timing. Since patients undergoing resection of hepatic metastases have been consistently shown to have improved survival [39], the response rate of a chosen regimen becomes relevant as it may be possible to convert a borderline resectable lesion to one where resection is possible. Adding an anti-EGFR agent in this setting can be considered given the higher response rates reported [40]. Additionally, as a consequence of improved therapies and longer survival, there is greater opportunity for surgical intervention in managing the complications of an in situ primary tumour or metastatic disease.

Most patients presenting with metastatic disease in colon cancer do not require urgent surgery. For those patients with malignant bowel obstruction, tumour perforation or bleeding, the value of urgent surgical intervention is not in doubt. Alternatively for symptomatic but well patients, chemotherapy has a 40–50 % chance of shrinking the primary tumour [1], which may improve symptoms of itself. Colonic stenting has also emerged as a feasible alternative that may avoid extensive surgery and improve obstructive symptoms [41, 42]. Stenting has the advantage that observational studies suggest chemotherapy may be administered promptly after stenting without apparent excess risk [43, 44].

There is considerable debate about the risks and benefits of elective resection of the primary tumour in asymptomatic patients with clearly incurable disease. The rationale for this approach is threefold. Firstly, it may prevent development of acute complications during the lifetime of the patient, which as noted previously is now significantly longer. In patients with metastatic colorectal cancer where the primary tumour remains in situ and there are no symptoms, only 11–14 % experience morbidity related to the primary tumour that may require surgical or nonsurgical intervention such as stenting or radiotherapy [45–48]. Secondly, primary tumour resection may prevent treatment complications such as haemorrhage or perforation that arise due to the use of the anti-VEGF agent bevacizumab. In a meta-analysis not incorporating individual patient data, bevacizumab was associated with an increased risk of gastrointestinal perforation in patients with metastatic colorectal carcinoma [49]. However, data pertaining specifically to primary tumour complication rates in patients with asymptomatic primary tumours receiving bevacizumab do not show increased rates of perforation or haemorrhage [47]. Thirdly, primary resection may improve the efficacy of systemic treatment and prolong survival. There are, however, conflicting data on the survival benefit of primary tumour resection and no prospective randomised trials [50], although recent retrospective studies have suggested a survival benefit [51, 52]. These data are tempered by the significant post-operative complication rates for primary tumour resection in patients with stage IV disease, with a 12 % major complication rate reported in one large retrospective study [45]. Retrospective studies have identified various factors that are associated with better survival after primary tumour resection including younger age, few comorbidities, liver-only metastases and lower serum carcinoembryonic antigen level [51–54]. Evidence from trials of neoadjuvant chemotherapy indicates that the primary tumour may be less sensitive to chemotherapy than metastatic lesions [55], but whether this difference is relevant to modern chemotherapy regimens with response rates of 40 % or more is unclear. In summary, it is likely that carefully selected patients may benefit from primary tumour resection followed by systemic therapy, but this patient population remains undefined.

As mentioned earlier with regard to surgical resection of hepatic metastases, metastectomy has a role in the management of advanced colorectal cancer and permits some patients with few sites of metastatic disease to achieve prolonged disease-free survival without chemotherapy. The most robust evidence exists for patients with metastases limited to the liver. In these patients who undergo resection of metastatic lesions, 5-year overall survival may average 40 %, with a median survival of 65 months [39, 56–58]. For patients considered borderline resectable or unresectable, there is a further subgroup that may be converted to resectable with systemic therapy [40, 59]. Improved outcomes are evident for patients converted to resectable and undergoing surgery, with a median survival of 40 months in one study, double the expected median survival for metastatic colorectal cancer [59]. This approach has not been validated in large randomised controlled trials, however. Less robust data also exist for the benefits of resection of limited pulmonary metastases [60, 61], although this consists mainly of retrospective analyses subject to bias and in theory randomised trials are needed [62].

Multi-agent chemotherapy is associated with improved outcomes in gastric cancer [63]. The most active regimens are considered to consist of capecitabine, a platinum compound such as cisplatin or oxaliplatin, and epirubicin, with response rates of around 40–48 % [64]. Trastuzumab in combination with capecitabine/5-fluorouracil and cisplatin has also shown efficacy in tumours that overexpress HER2 although this equates to only 20–25 % of patients [65]. More recently evidence from randomised phase III trials of chemotherapy with irinotecan or docetaxel versus best supportive care alone has shown the survival benefits of further treatment after the failure of first-line therapy [66, 67]. As with colorectal cancer, survival has now improved, and in particular for HER2-positive patients, median survival is reported as greater than 14–16 months [65], and thus the potential for late-stage events requiring surgery increases.

Pancreatic cancer is considered among the most lethal upper gastrointestinal malignancies. With treatment, median survival for locally advanced inoperable disease approaches 12 months in contemporary clinical trials [68] and 6–8 months for metastatic disease [69]. Chemoradiotherapy is the standard of care for patients with good performance status and inoperable disease, usually following several months of chemotherapy alone to determine if the tumour is responsive to chemotherapy. Compared to chemotherapy alone, chemoradiotherapy prolongs survival but may not improve quality of life [68]. Radiological response rates to chemoradiotherapy are less than 10 % [68]. For metastatic disease, single-agent gemcitabine is commonly used but suffers from poor response rates and an uncertain impact on quality of life [70, 71]. Recently a multi-agent regimen consisting of 5-fluorouracil, oxaliplatin and irinotecan (FOLFIRINOX) has produced the most promising results to date in metastatic pancreatic cancer, with a median survival of 11 months and response rate of 30 % [72]. FOLFIRINOX also prevented degradation in quality of life compared with gemcitabine [73]. Significant toxicity was associated with FOLFIRINOX, and it is considered suitable only for fit patients without troublesome biliary obstruction. That said, favouring surgical palliation of biliary obstruction to avoid the risk of infected biliary stents may become a more frequent discussion point to facilitate greater use of this regimen. Furthermore, the higher response rates seen with FOLFIRINOX warrant investigation in the preoperative setting with the hope that some patients may be down staged and become resectable.

Gastric outlet obstruction may develop later in the natural history of gastric cancer and pancreatic cancer or at presentation. It is typically associated with a decline in the patient’s functional status and nutritional state. In these unwell patients, the response rates expected with the available chemotherapy regimens for gastric, and in particular pancreatic, cancer are often inadequate to effectively palliate outlet obstruction. Although gastrectomy remains a successful intervention for gastric outlet obstruction related to gastric cancer, endoscopic stenting may be a preferable option for patients with more limited life expectancy or where surgery is not possible [74]. There are, however, no data on quality of life outcomes post-endoscopic stenting [75], and some authors have suggested the technology of duodenal stenting lags behind that of biliary stents, and thus laparoscopic or open gastric bypass remains an important consideration [76]. Radiation has been associated with good palliation in gastric cancer causing obstruction with symptom control rates of 80 % in a small series and has the advantage of also controlling bleeding [77]. Management of gastric outlet obstruction may then permit the administration of systemic chemotherapy.

The presence of uncontrolled biliary obstruction due to upper gastrointestinal malignancy or other disease metastatic to the liver renders systemic therapy problematic, affecting the excretion and toxicity profile of several agents active in gastrointestinal cancers such as epirubicin, gemcitabine, irinotecan and sorafenib [78]. Uncontrolled biliary obstruction also predisposes patients to cholangitis which is undesirable during potentially immunosuppressive chemotherapy and often excludes patients from inclusion in clinical trials. The alleviation of biliary obstruction is therefore preferable prior to instituting systemic therapy if possible. Stenting endoscopically or percutaneously is an effective and less invasive option than surgical bypass, having been shown to have fewer short-term complications, although a higher re-occlusion rate [79]. For patients with bulky metastatic disease, a borderline performance status or poorly controlled symptoms, stenting procedures allow prompt institution of systemic therapy without the need for recovery from surgery. In one small retrospective study in patients with metastatic colorectal or gastric cancer, chemotherapy administration was feasible in approximately half of patients who underwent successful percutaneous biliary drainage [80]. Factors associated with a poor outcome after the percutaneous procedure in this study included extensive liver metastases, poor performance status, prior chemotherapy and ascites [80]. Patients with a poor prognosis and limited life expectancy are also candidates for stenting for palliative benefit. Alternatively, in patients with low-volume disease and good functional status who are candidates for intensive chemotherapy such as FOLFIRINOX, the better long-term outcomes and theoretically lower risk of sepsis may favour surgical bypass, although stenting remains the preferred procedure for the majority of patients [79]. The multidisciplinary setting is the ideal environment to weigh these factors prior to intervention.