pPoliteness principles of Leech

Clinical Gastroenterology and Hepatology 2014;12:139–144

Personalizing Therapy for Colorectal Cancer Ashley Wong and Brigette B. Y. Ma Sir Y. K. Pao Centre for Cancer, Department of Clinical Oncology, Prince of Wales Hospital, Chinese University of Hong Kong, Hong Kong SAR, China Colorectal cancer (CRC) is the third most commonly diagnosed cancer worldwide. Several important scientific discoveries in the molecular biology of CRC have changed clinical practice in oncology. These included the comprehensive genome-wide profiling of CRC by the Cancer Genome Atlas Network, the discovery of mutations along the RAS-RAF signaling pathway as major determinants of response to antibodies against the epidermal growth factor receptor, the elucidation of new molecular subsets of CRC or gene signatures that may predict clinical outcome after adjuvant chemotherapy, and the innovative targeting of the family of vascular endothelial growth factor and receptors. These new data have allowed oncologists to individualize drug therapy on the basis of a patient’s tumor’s unique molecular profile, especially in the management of metastatic CRC. This review article will discuss the progress of personalized medicine in the contemporary management of CRC. Keywords: Colorectal Cancer; KRAS Mutation; Gene Signatures.

olorectal cancer (CRC) is the third most commonly diagnosed cancer in both men and women worldwide.1 The last decade has witnessed some important scientific discoveries in the molecular biology of CRC that have resulted in dramatic shifts in the treatment paradigms for metastatic CRC. These included the comprehensive genome-wide profiling of CRC by the Cancer Genome Atlas Network,2 the discovery of mutations along the RAS-RAF signaling pathway as major determinants of response to antibodies against the epidermal growth factor receptor (EGFR), the elucidation of new molecular subsets of CRC that may predict clinical outcome after adjuvant chemotherapy,3,4 and the innovative targeting of the family of vascular endothelial growth factor (VEGF) and receptors (VEGFRs). This body of knowledge has not just enabled us to individualize drug treatment on the basis of molecular profiles; they have contributed to the improvement of prognosis in advanced CRC. The latter has been achieved through the development of new targeted therapies and by changing the way we manage patients with oligometastases from CRC. This review article will discuss the progress of personalized medicine in the contemporary management of CRC.

C

Personalizing Drug Therapy for Colorectal Cancer: An Evolving Concept The concept of individualized drug therapy in CRC is not new, and traditional factors such as a patient’s performance status, presence of medical comorbidities, TNM staging, and specific histopathologic characteristics still heavily influence the decision-making process of oncologists. However, this approach is limited because it does not adequately explain the differential response to drug therapy in individuals; furthermore, it may inadvertently result in the overtreatment of patients. The latter is exemplified by the adjuvant treatment of stage II CRC, where the number of patients needed to be treated with adjuvant chemotherapy to prevent 1 recurrence or death is 25–50 patients, at the expense of 1 in 6 patients experiencing serious toxicity.5 Genome-wide molecular profiling studies have shown that CRC is a heterogeneous disease characterized by multiple genetic and epigenetic alterations. Molecular changes such as microsatellite instability (MSI), CpG island methylator phenotype (CIMP), global DNA hypomethylation, and chromosomal instability that result in the activation of oncogenic pathways are commonly found in most CRC tumors.6 The reader should refer to some excellent reviews on the molecular pathogenesis of CRC, because a detailed discussion on this topic is beyond the scope of this review.6–9 Recently, the landmark publication by the Cancer Genome Atlas Network on CRC has reported new driver mutations and provided a unifying view of the known genetic and epigenetic aberrations.2 This seminal work has shown that colon and rectal cancers are similar on a genomic level, whereas proximally located primary tumors are more likely to exhibit a “hypermutated” genomic profile than distally located tumors. This profile consists of much higher frequencies of somatic mutations, microsatellite unstable status,

Abbreviations used in this paper: CI, confidence interval; CIMP, CpG island methylator phenotype; CRC, colorectal cancer; EGFR, epidermal growth factor receptor; 5-FU, 5-fluorouracil; HR, hazard ratio; MSI, microsatellite instability; PFS, progression-free survival; OS, overall survival; PI3K, phosphoinositide 3-kinase; VEGF, vascular endothelial growth factor; VEGFR, vascular endothelial growth factor receptor. © 2014 by the AGA Institute 1542-3565/$36.00 http://dx.doi.org/10.1016/j.cgh.2013.08.040

140

Wong and Ma

mismatch repair gene mutation, and CIMP. Apart from confirming that genetic alterations are common in the Wingless (Wnt), phosphoinositide 3-kinase (PI3K)-AKT, RAS-RAF-ERK, and transforming growth factor receptorbeta signaling pathways, newer aberrations such as amplification of HER2 and insulin-like growth factor 2 and somatic mutations of AT-rich interactive domaincontaining protein 1A, sex determining region Y-box 9, and family with sequence similarity 123B may represent potential biomarker or therapeutic targets.2 In the following sections, the translational significance of these basic scientific discoveries will be presented in 2 clinical settings, metastatic and adjuvant.

Personalizing Drug Therapy in the Metastatic Setting Mutations Along the RAS-RAF and Phosphoinositide 3-Kinase Signaling Pathway Cetuximab and panitumumab are monoclonal antibodies targeting the extracellular domain of the EGFR and have been used in clinics for more than a decade. Historically, these drugs were once administered in unselected patient population, with modest gains in response rate and progression-free survival (PFS).10–15 The RAS-RAF signaling is one of the effector pathways located downstream to the EGFR, which also cross-talk with the PI3K-AKT pathway. Activating mutations of KRAS, BRAF, and PIK3CA have been implicated as escape mechanisms that can bypass the growth inhibitory effect of EGFR blockade.16 KRAS mutations can be found in around 40% of CRC, of which around 90% occur at codon 12 or 13.16 Since mid-2005, data began to emerge

Clinical Gastroenterology and Hepatology Vol. 12, No. 1

on the strong association between KRAS mutation and resistance to EGFR antibodies in retrospective series and post hoc analyses of phase III studies (Figure 1).17–19 Collectively, these studies have shown that patients with KRAS mutant cancers do not derive any added benefit in terms of tumor control and survival when being treated with EGFR antibodies alone or in combination with chemotherapy, perhaps with the exception of some patients with the relatively uncommon G13D mutation.20 The drug insert for cetuximab and panitumumab has since been revised, recommending mandatory testing for KRAS mutation before starting therapy. To cite an example of the powerful predictive effect of KRAS mutation on clinical response to anti-EGFR antibodies, the phase III CRYSTAL study was a pivotal study that led to the approval of cetuximab in combination with chemotherapy in the first-line treatment of metastatic CRC.10 The hazard ratio (HR) for the primary end point of PFS among patients treated with cetuximab plus chemotherapy who had tumors with wild-type KRAS was 0.68 (P ¼ .02), compared with those with KRAS-mutant tumors, HR ¼ 1.07 (P ¼ .75). Patients with KRAS wildtype tumors also had a higher response rate of 59.3% than those with KRAS-mutant tumors (36.2%) after treatment with cetuximab chemotherapy.10 Besides KRAS mutation, accumulating evidence suggests that mutations in other downstream components of EGFR pathway affecting the BRAF, PIK3CA, NRAS, and PTEN genes have also been associated with poor prognosis after treatment with anti-EGFR antibodies in metastatic CRC.21–23 However, clinical application of these biomarkers is limited by their relatively low level of expression in CRC. To date, no oncological guidelines have recommended routine testing for these markers in metastatic CRC.

Figure 1. Selected signaling pathways of clinical significance in CRC and therapeutic targets. Schematic diagram of approved targeted agents for treatment of advanced CRC and their mode of action in a CRC cell. AKT, protein kinase B; BAD, Bcl-2-associated death promoter; MAPK, mitogenactivated protein kinase; mTOR, mammalian target of rapamycin; NFkB, nuclear factor k-light-chainenhancer of activated B cells; PIGF, phosphatidylinositol-glycan biosynthesis class F protein.

January 2014

Cancer Pharmacogenomics Cancer pharmacogenomics is also another area that has been extensively investigated in CRC. Certain genotypic variants such as those affecting the genes encoding the UDP-glucuronosyltransferase 1A1 enzyme (which is involved in the metabolism of irinotecan) and dihydropyrimidine dehydrogenase (involved in the metabolism of 5-fluorouracil [5-FU]) are now available for clinical testing. However, the cause of an individual’s susceptibility to drug toxicity is usually multifactorial involving multiple genetic as well as non-genetic (eg, drug scheduling) factors. For instance, the UGT1A1 polymorphism is not as predictive of severe toxicity to irinotecan in patients receiving the more popularly used, 2-weekly schedule of irinotecan than the less commonly used 3-weekly schedule.24 Common polymorphisms in genes encoding for other proteins such as adenosine triphosphate binding cassette and solute carrier transporters may also contribute to increased vulnerability to irinotecan toxicity.25 Therefore, the clinical impact and applicability of these tests remain controversial and are not in routine use in Hong Kong and most Asian centers.26

Personalizing the Systemic Treatment of Oligometastases For the management of patients with metastases that are limited to the liver or lung, the aim is to optimize tumor shrinkage with more intensive drug regimens to achieve potentially curative resection. The availability of more effective neoadjuvant drug therapy has dramatically changed the prognosis of patients with initially inoperable or borderline resectable liver metastases. The 5-year survival of patients who presented with inoperable liver metastases is now approaching that of patients who presented with operable tumors, after multimodal treatment with chemotherapy, liver resection, and local ablative therapy.27 The addition of targeted therapy such as anti-EGFR antibodies to chemotherapy in patients with KRAS–wild-type liver metastases has further extended tumor response rate to greater than 70%, R0 resection rate of greater than 30%, and also overall survival (OS) compared with chemotherapy alone in a recently published study.28–30 In a phase III study reported by Ye et al,30 patients with KRAS wild-type CRC who were randomized to chemotherapy and cetuximab experienced higher 3-year OS rate (41% vs 18%, P ¼ .013) and median OS (30.9 vs 21.0 months, P ¼ .013), compared with those who received chemotherapy alone. In recognition of the improved prognosis of patients with oligometastases, the latest American Joint Committee on Cancer classification of CRC has divided the M stage into M1a (oligometastases) and M1b (multiple sites of metastases).31

Colorectal Cancer

141

Targeting Tumor Angiogenesis in Metastatic Colorectal Cancer VEGF is the key regulator of angiogenesis that is essential for tumor growth and metastasis in CRC. Bevacizumab is the first anti-angiogenesis drug that has been approved in the treatment of metastatic CRC (and also for any cancer) in 2004.32 Two recently reported phase III studies may have practice-changing implications. The first of these 2 studies is the TML study, which is the first phase III study to show that in patients with progressive metastatic CRC during first-line treatment with bevacizumab and chemotherapy, the continuation of VEGF inhibition with bevacizumab when switching across to subsequent line of chemotherapy will confer added survival benefit. Patients who were continued on bevacizumab on subsequent line of chemotherapy had higher median OS (11.2 months) than those who did not (9.8 months; HR, 0.81; 95% confidence interval [CI], 0.69–0.94; P ¼ .0062), compared with if bevacizumab is discontinued.33 The second study, dubbed the DREAM study, is one of the first reported phase III studies that investigated the use of bevacizumab as a maintenance therapy in combination with erlotinib (an EGFR tyrosine kinase inhibitor), after optimal response to first-line chemotherapy for metastatic CRC.34 The DREAM study showed that the bevacizumab-erlotinib arm resulted in a longer median PFS of 5.8 months vs 4.6 months in the bevacizumab alone arm (HR, 0.73; 95% CI, 0.59–0.91; P ¼ .005). Both the TML and DREAM studies enrolled patients who were medically fit and excluded those with rapidly progressive disease or poor response to chemotherapy. Therefore, careful selection of patients for these 2 different treatment strategies is needed in practice. Two new drugs targeting angiogenesis have recently been approved for the treatment of metastatic CRC. Aflibercept is a fully human recombinant fusion protein that consists of the VEGF-binding portions from the extracellular domains of human VEGFR 1 and 2 and the Fc portion of human immunoglobulin G1. It acts as a decoy by binding to the circulating angiogenic factors VEGF-A, VEGF-B, and placental growth factor, so that they cannot bind to the respective receptors.35 In the registrational, phase III VELOUR study, the addition of aflibercept to chemotherapy significantly improved OS and PFS in patients receiving second-line chemotherapy. Patients who received aflibercept chemotherapy had better PFS (6.90 vs 4.67 months; HR, 0.75; 95% CI, 0.661–0.869; P