The advent of affordable technologies to perform detailed molecular profiling of tumours has transformed understanding of the specific genetic events that promote carcinogenesis and which may be exploited therapeutically. The application of targeted therapeutics has led to improved outcomes in advanced disease and this approach is beginning to become established in the management of potentially curable disease for surgical patients.
This review article focuses on recent developments in the management of operable cancers of the gastrointestinal (GI) tract, specifically discussing the currently available data that evaluate the incorporation of targeted therapies in this setting.
A variety of targeted molecules are now available as treatment options in the management of GI cancers. Most are aimed at growth inhibition by acting on cell surface targets or intracellular pathways. Treatment paradigms are gradually shifting towards more prevalent use of systemic treatment prior to surgical intervention for operable disease with the aim of tumour downsizing and improved rates of long-term cure.
Over the past 15–20 years there have been unprecedented advances in the field of molecular biology and associated technologies that allow genomic analysis of tissues. These advances have allowed the oncology community to advance greatly its understanding of cancer pathogenesis. It is now understood that individual tumours are frequently driven by the acquisition of key genetic abnormalities known as ‘oncogenic drivers’ that individually or in combination promote a number of behavioural features key to the survival of the cancer cells. These drivers may promote a more aggressive phenotype through increased cell proliferation and greater potential for invasion and metastasis, or permit the cancer cell to evade apoptotic signals or survive the cytotoxic insult from traditional oncological therapies. Similarly, aberrations leading to loss of function of host tumour suppressor genes may further facilitate tumorigenesis.
Alongside this knowledge of the molecular abnormalities driving tumour formation and growth, considerable research has been undertaken to investigate the mechanisms by which these processes may be targeted. Many of the currently available compounds target specific growth factor receptors expressed on the cancer cell surface or attempt to block the growth stimulatory pathways within the cell. There has been a resultant explosion of novel ‘targeted therapies’, mainly in the form of either monoclonal antibodies that bind to the extracellular domain of a key receptor, or small-molecule inhibitors that bind to tyrosine residues on the intracellular domain of the receptor and block the growth stimulatory signals that occur through the tyrosine kinase pathway. The aim of these therapies is primarily to reduce signalling via oncogenic intracellular pathways and thereby alter the transcriptional events that occur within the cancer cell nucleus (Fig.1).
These therapies may have potential benefits in all tumour types and at all stages in the cancer process. Success depends not on the tissue of origin or histological subtype—the factors used to make traditional chemotherapy decisions—but rather on the degree of addiction of an individual tumour to the pathway being targeted. Importantly, some of these targeted drugs have now demonstrated meaningful benefits in the neoadjuvant or perioperative settings where they may influence surgical practices, for instance by improving resectability rates as well as increasing the number of patients cured. This article aims to discuss the use of targeted therapies in the neoadjuvant setting and their influence on the treatment approach for potentially operable tumours of the gastrointestinal (GI) tract. As so many important lessons have been learned from cancers outside of the GI tract, notably breast and kidney, these have been included and provide models that may influence the management of GI tumours in the future.
Rationale for neoadjuvant therapy for primary tumours
Treatment given in the neoadjuvant setting for potentially operable disease has a number of potential benefits including tumour downsizing, increased likelihood of complete macroscopic and microscopic clearance of the primary tumour (R0 resection), earlier treatment of micrometastatic disease and the ability to assess responsiveness to therapy directly. Conversely, it can also be a cause for some concern owing to the delay to surgery with the risk of disease becoming inoperable, the possibility of clinical deterioration and the potential for increased perioperative complications. It is therefore important to demonstrate that the number of patients who benefit from this approach exceeds the number harmed. Neoadjuvant strategies are perhaps best established in the treatment of breast and rectal cancers where they form current standards of care. However, with the advent of more effective targeted therapies, the use of preoperative treatment is being explored in a number of different solid organ tumours.
Targeted neoadjuvant therapies—proof of efficacy from breast cancer
The demonstration of hormone (oestrogen, progesterone) receptors and epidermal growth factor family receptors on tumour cell surfaces in patients with breast cancer has been recognized for over 30 years, along with their influences on survival. Attempts to target these receptors or pathways under their control represent some of the earliest efforts at a targeted approach to cancer treatment.
Human epidermal growth factor receptor 2 (HER-2) positive breast cancer is currently characterized as being either strongly positive on immunohistochemistry (IHC 3+) or moderately positive (IHC 2+) but with confirmation of expression by fluorescence in situ hybridization. The randomized phase III NOAH trial (235 patients) evaluated the addition of trastuzumab (an anti-HER-2 monoclonal antibody) to neoadjuvant chemotherapy (doxorubicin, paclitaxel, cyclophosphamide, methotrexate and fluorouracil)1. Patients who were randomized to the trastuzumab arm completed a total of 1 year of therapy (neoadjuvant plus adjuvant). Addition of trastuzumab led to significant improvements in pathological complete response (pCR) rate (43 versus 22 per cent; P < 0·001) and 3-year event-free survival rate (71 versus 56 per cent; hazard ratio 0·59; P = 0·013). Following these findings, trastuzumab has become incorporated routinely into the neoadjuvant and adjuvant treatment of HER-2-positive breast cancers.
These promising results with trastuzumab led to the evaluation of similar agents in the neoadjuvant setting, including lapatinib (a tyrosine kinase inhibitor (TKI) targeting both EGFR and HER-2) and pertuzumab (a monoclonal antibody that prevents dimerization of the HER-2 receptor with three other HER family receptors). In the randomized phase III Neo-Adjuvant Lapatinib and/or Trastuzumab Treatment Optimization (NeoALTTO) trial (455 patients), HER-2-positive patients received neoadjuvant chemotherapy with paclitaxel plus either lapatinib, trastuzumab or both (dual anti-HER-2 strategy)2. The highest pCR rates were achieved with the dual strategy compared with either lapatinib or trastuzumab alone (51·3 versus 24·7 versus 29·5 per cent respectively; P < 0·001). The Neoadjuvant Study of Pertuzumab and Herceptin in an Early Regimen Evaluation (neoSPHERE) study evaluated pertuzumab, in combination with docetaxel and/or trastuzumab3. Patients treated with all three drugs (107 patients) achieved a pCR rate of 45·8 per cent; however, perhaps most strikingly, there was a 16·8 per cent pCR rate with the combination of pertuzumab plus trastuzumab (107 patients), confirming that targeted therapy combinations can demonstrate significant efficacy in the absence of cytotoxic chemotherapy.
In HER-2-negative breast cancer, other targeted therapies might have a role in the neoadjuvant setting, although the current results are conflicting. Bevacizumab is a humanized monoclonal antibody that inhibits vascular endothelial growth factor (VEGF) A. Both the GeparQuinto (1948 patients) and National Surgical Adjuvant Breast and Bowel Project (NSABP) B-40 (1206 patients) trials confirmed increased pCR rates with the addition of bevacizumab to chemotherapy4, 5. In the GeparQuinto study, however, benefit was confined largely to the triple-negative population (oestrogen receptor-, progesterone receptor- and HER-2-negative) where the addition of bevacizumab to epirubicin, cyclophosphamide and docetaxel improved the pCR rate from 27·9 to 39·3 per cent (P = 0·003)4. There was no benefit associated with bevacizumab in the 1262 patients with hormone receptor-positive tumours. In contrast, the NSABP B-40 study found bevacizumab to improve pCR rates significantly among hormone receptor-positive patients (23 versus 15 per cent), with no significant effect in the triple-negative group5. Given the inconsistency of these results, the lack of predictive biomarkers and the lack of long-term outcome data, the use of bevacizumab in this setting is not currently recommended.
Lessons learned from renal cell carcinoma
Renal cell carcinoma (RCC) represents a relatively common cancer in which molecularly targeted therapies have led to dramatic successes where cytotoxic chemotherapies had previously failed. Several small-molecule protein kinase inhibitors that block angiogenesis, Akt/mTOR and/or Ras/Raf pathways (including sunitinib, sorafenib, pazopanib and everolimus) now have confirmed efficacy in the setting of advanced disease. In early-stage operable disease there is an increasing tendency towards partial rather than radical nephrectomy where possible. Partial nephrectomy represents a safe option in terms of cancer control and has the benefit of preserving long-term renal function6, 7, which may also improve long-term overall survival by lowering the risk of cardiovascular disease7. These findings have led to interest in the use of neoadjuvant TKI therapy to downsize larger tumours and allow more patients to undergo nephron-sparing surgery.
In a prospective clinical trial of preoperative sunitinib, some degree of tumour shrinkage was seen in 17 of 20 patients, with no apparent increase in surgical complications8. Case reports suggest that, for patients who need to avoid radical nephrectomy, it is possible to achieve adequate downsizing with sunitinib to allow partial nephrectomy9, 10. Preoperative sorafenib has also been investigated in 30 patients before nephrectomy, and was associated with a decrease in primary tumour size and loss of intratumoral enhancement in the majority of patients (83 and 88 per cent respectively) with no surgical complications related to therapy11. A series evaluating both sunitinib and sorafenib found that both were associated with increased incidence and severity of intraoperative adhesions that made surgery more difficult, although this did not translate into an increase in perioperative complications12. It is not difficult to see how this principle might be applied in the GI tract, for instance in the context of neoadjuvant treatment before liver surgery, where use of targeted therapies may avoid some of the steatohepatitic changes seen with conventional chemotherapies.
In operable oesophagogastric adenocarcinoma, the use of perioperative chemotherapy represents a standard of care following confirmed overall survival benefit in randomized phase III trials13, 14. At the present time, targeted therapies do not form part of the treatment regimen in this setting, although biomarkers of interest are now beginning to emerge. In HER-2-positive gastric cancer, following the success of the Trastuzumab for GAstric cancer (ToGA) trial in advanced disease15, there are now several case reports of patients receiving chemotherapy plus trastuzumab in the perioperative setting, with promising results16, 17. Current prospective trials are therefore evaluating trastuzumab and other anti-HER-2 therapies within perioperative regimens in HER-2-positive disease. Additionally, in the UK the ST03 trial is currently evaluating the addition of bevacizumab to chemotherapy in patients with operable adenocarcinoma of the stomach and oesophagus, and is undergoing amendment to include perioperative lapatinib in the HER-2-positive subset18.
In oesophageal carcinoma, cetuximab (a chimeric mouse–human epidermal growth factor receptor (EGFR) inhibitor) has been evaluated in combination with neoadjuvant chemoradiotherapy within a phase II trial19. Among 28 patients with adenocarcinoma or squamous cell carcinoma of the oesophagus, 19 achieved a complete or ‘near complete’ pathological regression, and treatment toxicities were considered acceptable with no apparent increase in surgical complications and no treatment-related deaths. Currently, therapies targeting the MET and fibroblast growth factor receptor pathways are undergoing evaluation in advanced oesophageal cancer, and successful therapies can be expected subsequently to enter trials in the operable setting.
Gastrointestinal stromal tumours
The impact of targeted therapies within the past decade has been more evident in gastrointestinal stromal tumours (GISTs) than in any other GI cancer. GIST previously represented a disease entity that was poorly classified, largely refractory to standard chemotherapies, and where surgical excision represented the only effective intervention. Following the discovery that these tumours were nearly entirely driven by activating mutations in the c-KIT (over 90 per cent)20, 21 and platelet-derived growth factor receptor α (PDGFR-A) proto-oncogenes, a targeted therapeutic approach has proven highly successful. Imatinib is a small-molecule tyrosine kinase inhibitor that effectively blocks both KIT and PDGFR-A signalling. In advanced GIST imatinib has been demonstrated to improve median overall survival dramatically from approximately 20 months to nearly 5 years22. However, not all activating mutations confer the same degree of sensitivity to imatinib therapy. KIT exon 11 mutations are associated with increased benefit compared with exon 9 mutations or where there is no identifiable mutation23–25. For this reason mutational analysis is recommended for all patients, and use of an increased starting dose of imatinib is advocated in those with exon 9 mutations26.
Given the impressive response rates with imatinib in advanced disease, there has been considerable interest in the potential benefits of this therapy prior to resection in an attempt to improve long-term outcomes. This is a particularly attractive strategy in patients with unresectable primary tumours, or where surgery would result in multivisceral resections with significant associated morbidity. It may also be of benefit in the treatment of local recurrences or in the management of potentially resectable but small-volume metastatic disease. The use of neoadjuvant imatinib has never been evaluated in a prospective randomized clinical trial, although a rationale for its use is supported by case reports, observational studies and a single-arm phase II study in patients with either locally advanced primaries (30 patients) or resectable recurrent/oligometastatic disease (22 patients)27. All patients received 8–12 weeks of neoadjuvant imatinib followed by at least 2 years of adjuvant therapy. In the locally advanced patient group, 57 per cent remained progression-free at 5 years, and even in those who were recruited with recurrent or metastatic disease the 5-year progression-free survival rate was 30 per cent.
In a reported retrospective series of 46 patients treated with neoadjuvant imatinib followed by surgery, there were 11 patients with locally advanced tumours28. Median duration of imatinib therapy in this group was 12 months, and one patient achieved a pCR with a further eight achieving a radiological partial response. Only one of the 11 patients experienced disease recurrence after nearly 2 years following resection. Despite the lack of randomized data, neoadjuvant imatinib is often recommended in those with high-risk features including: size greater than 5 cm, more than five mitoses per 50 high-power fields, or anatomical location that would necessitate a potentially morbid resection. The European Society for Medical Oncology guidelines state that neoadjuvant imatinib is recommended if ‘R0 surgery is not feasible, or it could be achieved through less mutilating surgery in the case of cytoreduction’26. This advice is echoed by that of the National Comprehensive Cancer Network in America29.
Colorectal cancers also have an interesting history in the evolution of targeted therapies. One of the intracellular growth signalling pathways activated by EGFR is the mitogen-activated protein kinase (MAPK) pathway, involving a KRAS protein. About 40 per cent of patients with colorectal cancer have tumours with mutant KRAS. When KRAS is mutated, it does not respond to the usual EGFR-mediated signals and unrestricted growth can occur. Attempts to use EGFR inhibitors in patients with advanced colorectal cancer have therefore demonstrated that this treatment only improves outcomes in patients with wildtype (non-mutated) KRAS. This in turn led to the development of genetic testing for KRAS mutation as a predictive biomarker of response to treatment. This was the first genetic test to be used to guide cancer treatment.
Colon cancer is largely resectable at presentation and therefore neoadjuvant therapy is not routinely required. However, for locally advanced tumours, the cancer may be fixed to surrounding structures and therefore neoadjuvant therapy may improve operability. Targeted therapies have not been explored in this setting previously; however, this is one of the goals of the current Fluoropyrimidine Oxaliplatin and Targeted Receptor Pre-Operative Therapy (FOxTROT) trial18. This trial is designed to evaluate whether neoadjuvant therapy in stage T3/T4 colorectal cancer will improve disease-free survival, and whether the addition of panitumumab (a human monoclonal antibody that specifically blocks EGFR) to chemotherapy will increase tumour shrinkage in patients with KRAS wild-type tumours.
In rectal cancer, neoadjuvant therapy is commonly recommended to achieve local tumour shrinkage and facilitate total mesorectal excision for bulky tumours, where invasion of adjacent organs or the walls of the pelvis can compromise the ability of the surgeon to achieve an R0 resection. In addition, it is desirable to avoid the permanent colostomy and high incidence of urinary or sexual dysfunction associated with abdominoperineal resection for lower rectal tumours, by achieving sufficient tumour regression to permit a satisfactory (R0) resection while restoring intestinal continuity and retaining sphincter continence. Neoadjuvant chemoradiotherapy therefore represents the current standard of care for patients with magnetic resonance imaging-defined high-risk features, including a potentially positive circumferential resection margin (CRM), extramural venous invasion and extramural spread beyond 5 mm. The goal of therapy is to downsize the tumour sufficiently to allow a sphincter-sparing operation, and decrease the chances of an involved CRM.
Following data confirming a radiosensitizing role for cetuximab in head and neck cancers30, the randomized phase II EXPERT-C trial evaluated the addition of cetuximab to a schedule of neoadjuvant chemotherapy, followed by chemoradiotherapy, surgery and adjuvant capecitabine/oxaliplatin for rectal cancer31. This trial recruited 90 patients with KRAS/BRAF wild-type tumours and the primary endpoint of the study was pCR rate. Although the addition of cetuximab did not improve pCR rates, there were statistically significant improvements in both the response to neoadjuvant chemotherapy (71 versus 51 per cent; P = 0·038) and overall survival (hazard ratio 0·27; P = 0·034) associated with the use of cetuximab in this setting.
Sorafenib is a TKI which acts to inhibit a number of different protein kinases, notably vascular endothelial growth factor receptor (VEGFR), PDGFR, and Raf. In 2008 the SHARP trial confirmed an overall survival benefit associated with the use of sorafenib compared with placebo in patients with advanced hepatocellular carcinoma (HCC) (median overall survival 10·7 versus 7·9 months; hazard ratio 0·69; P < 0·001)32. In early-stage disease amenable to surgery, partial liver resection represents the optimal surgical approach but this requires adequate functional liver reserve. Owing to underlying cirrhosis and liver dysfunction in the majority of patients with HCC, an appropriate oncological resection may not be feasible. Some patients harbouring small tumours may be candidates for liver transplantation. The potential use of neoadjuvant sorafenib as a holding measure before transplantation, and as a means of downsizing inoperable tumours to allow resection, has attracted considerable interest. Small case series have confirmed that in locally advanced disease it is possible to obtain meaningful improvements with sorafenib therapy that may subsequently render a tumour operable33, 34. Sorafenib can cause significant tumour necrosis and in some reports no viable tumour cells have been found in the resection specimen, although this needs to be tempered by concerns over its safety. Among 33 patients with HCC awaiting liver transplantation, ten received sorafenib in an attempt to prevent disease progression35. Although rates of death were similar between the two groups, the authors reported that the preoperative use of sorafenib was associated with statistically significant increases in the risk of biliary complications and acute cellular rejection following transplantation. Use of neoadjuvant sorafenib therefore remains experimental and requires further exploration within a clinical trial setting.
Owing to the anatomical location and aggressive nature of pancreatic cancers, these tumours are frequently inoperable at presentation, with less than one-fifth of patients presenting with operable disease. However, a further 15–20 per cent have locally confined disease which is inoperable by virtue of close proximity to, or direct invasion into, surrounding structures. There is therefore a potential role for treatment to downsize locally advanced tumours and render them operable. To date, targeted therapies have not established themselves in this setting, and current neoadjuvant treatment consists largely of either gemcitabine-based chemotherapy, chemoradiotherapy or both. Data on the use of triplet chemotherapy with oxaliplatin, irinotecan, fluorouracil and leucovorin (FOLFIRINOX) in this indication is currently awaited following confirmation of a 31·6 per cent response rate and more than 4-month improvement in overall survival in patients with metastatic disease36.
There is, however, increasing information regarding potential therapeutic targets for biological therapy in pancreatic cancer. KRAS mutations are found in 70–80 per cent of these tumours37, 38 and EGFR amplification is reported to occur in almost 50 per cent37. Further molecular classification has delineated three distinct subgroups of pancreatic cancer that may have differential responses to treatment39 and future studies may stratify according to these molecular signatures. Regarding targeted therapies, there have already been trials evaluating the use of cetuximab40, erlotinib41, 42 (another EGFR inhibitor), bevacizumab43–45 and axitinib46 (a TKI that inhibits multiple targets including VEGFR receptors, PDGFR and c-KIT) in combinations with chemotherapy in advanced disease. Of these, only the addition of erlotinib to gemcitabine has resulted in any improvement in overall survival41, the benefit being marginal. A retrospective series has suggested that the benefits of erlotinib may be predominantly limited to KRAS wild-type tumours47. The hedgehog signalling pathway is also currently being targeted in pancreatic cancer trials and the next decade is likely to see the development of more effective targeted therapies in this disease designed to improve resectability and long-term survival.
Neoadjuvant therapy for metastatic disease
In addition to the use of targeted therapies in the neoadjuvant setting, the advent of more effective systemic treatments has led to greater interest in surgical interventions for metastatic disease. As the overall survival of patients with advanced disease gradually improves, the potential role of palliative surgery alongside targeted therapies has been increasingly evaluated.
In colorectal cancer the addition of targeted therapies to improve treatment response and resectability of liver metastases has been explored widely within phase II trials, including the BOXER (bevacizumab)48 and CELIM (cetuximab)49 trials. This has also been evaluated and reported as a secondary endpoint within large phase III studies evaluating bevacizumab, cetuximab and panitumumab in advanced colorectal cancer50–53. In diseases such as GIST, RCC and melanoma, it is becoming increasingly common to resect known sites of metastatic disease following demonstration of response or stabilization with targeted therapy54–58. Additionally, case reports have described patients developing progressive disease at a single site who undergo resection of this one metastasis while continuing targeted therapy for control at other sites59. Since this phenomenon is likely due to development of resistant clones at an individual site of disease, surgery may have an increasing role in maintaining longterm disease control.
The number of targeted therapies being developed and evaluated in all stages of oncological clinical trials has increased dramatically in the past few years. These therapies are now impacting on the management of all stages of disease, including in the operable setting. Table1 lists some of the ongoing clinical trials evaluating a targeted therapy in the neoadjuvant setting in the GI tract. Patients who would previously have undergone surgery alone are increasingly receiving neoadjuvant therapies to improve the chances of a successful surgical outcome and long-term cure.
Table 1. Currently registered, recruiting trials in gastrointestinal cancer that are evaluating targeted therapies within the neoadjuvant setting
From ClinicalTrials.gov18. GOJ, gastro-oesophageal junction; HER, human epidermal growth factor receptor; pCR, pathological complete response.
Stomach + type III GOJ: stage Ib–IV, M0; type I/II GOJ, lower oesophagus: stage II–IVa
Colonic, wild-type KRAS; T4 or T3 with poor prognosis
Recurrence rate, pathological response
NCT00647530 (FOxTROT trial)
Rectal, wild-type KRAS; T3a–c
Rectal cancer; T3/4 or N1–2
Rectal cancer; ‘locally advanced’
Rectal cancer; T3/4 or N1–2
Toxicity, pathological response
Index tumour measuring 4–8 cm; Child–Pugh A
Radiological complete response, time to recurrence
Eligible for curative resection
Pathological response, antiangiogenic effects
LDE-225 (hedgehog inhibitor)
Safety, resection rate
GDC-0449 (hedgehog inhibitor)
CP-870,893 (CD40 agonist)
This process remains in its infancy and only a few molecularly targeted therapies to date have established themselves as components of routine clinical practice in early-stage disease. This will undoubtedly change over the next decade as understanding of carcinogenesis and resistance mechanisms continues to expand and new targets can be exploited. A wealth of biomarker data further aims to identify the specific subgroups that maximally benefit from each therapy, and effective biological therapies will increasingly be combined without the need for accompanying cytotoxic therapy. The increased use of techniques such as gene expression profiling, and reduced costs of deep sequencing technologies, are now permitting the complete genetic analysis of individual tumours. This approach is continuing to identify novel molecular aberrations associated with carcinogenesis, some of which may represent the therapeutic targets of the future.
This molecular approach to cancer treatment is therefore advancing towards a foreseeable situation where each patient would have their tumour profiled at presentation in order to determine the optimal combination of targeted therapies that would exploit the dependencies of their cancer. Throughout treatment, further analyses would then aim to identify emerging resistance mechanisms and allow a change in treatment strategy accordingly. As knowledge and treatment strategies continue to advance towards this goal, an increasing number of locally advanced tumours will be rendered resectable. Furthermore, as incurable cancers become increasingly chronic in their disease course, the resection of metastatic disease sites may become increasingly important as a means of prolonging overall survival.