Answer: When we talk about biomarkers in treatment options for colorectal cancer, one thing that really stands out are KRAS mutations as a negative predictive marker for epidermal growth factor (EGFR) antibodies, cetuximab and panitumumab. This is probably the highlight of a predictive biomarker: if the patient has the KRAS mutation, no treatment; if they have a wild type EGFR mutation, the antibodies have a chance to work.

This assumption was recently challenged with data presented at ASCO 2011, and again counterpointed at the European Society for Medical Oncology (ESMO) 2011 European Multidisciplinary Cancer Congress (EMCC) in Stockholm. KRAS mutations can occur at various codons. About 80% of all mutations are in codon 12. About 10%-12% of mutations are in codon 13, and there are some other mutations in codon 61 and other areas. What we saw in the presentation at ASCO this year confirmed the findings of an earlier paper, indicating that patients with a so-called G13D mutation, meaning codon 13 mutation, might actually still benefit from cetuximab. (Table 1) These were pooled retrospective analyses of cetuximab trials CRYSTAL and OPUS.^1,2 The findings implied that not all KRAS mutations are the same, and that cetuximab plus chemotherapy might benefit patients with this particular KRAS mutation. 

Data presented at the EMCC looked at pooled analyses of trials with panitumumab.³ In data from more than 1,400 patients, the G13D mutation was absolutely not associated with improved outcome with panitumumab. There was no consistent difference in benefit across trials in terms of progression-free survival (PFS) and overall survival (OS) for patients having the most common codon 12 and 13 KRAS mutant alleles.

How are we to interpret these data? Either the G13D mutation matters for cetuximab but not for panitumumab, which I don't think is true, or we can just say that cetuximab analysis is just a fluke. I think right now we are still back to the point that KRAS mutations are clearly negative predictive markers for panitumumab and I would also say for cetuximab. So I would not use EGFR antibodies in any KRAS-mutated tumors.

Now the other biomarker that I think is interesting in this context is BRAF. BRAF is in the signal transduction pathway just one step below KRAS. It might seem that BRAF mutations would also result in resistance to cetuximab and panitumumab, but what's really pronounced in BRAF-mutated tumors is the very poor prognosis. So BRAF could very well be a negative predictive marker but it's definitely a very strong poor prognostic marker – patients with BRAF mutated tumors live half as long as patients with non-BRAF mutated tumor, so this is a very strong biologic effect. It is very interesting to see that these molecules are only one step apart in our diagrams, but have very different biologic consequences when they are activated.

Beyond KRAS and BRAF, there are other interesting markers, particularly in terms of treatment with EGFR antibodies such as cetuximab and panitumumab. NRAS for instance is a parallel protein for KRAS and we know that the NRAS mutation, although found in just 1%-2% of patients, also serves as a negative predictive marker for cetuximab and panitumumab. This was highlighted in some of the analysis from PICCOLO trial, which used panitumumab in the second-line setting in combination with irinotecan.⁴ Patients who had any mutations in the KRAS, NRAS, or BRAF did not benefit from these antibodies.

I think in the future we won’t just focus on KRAS, we’ll integrate BRAF and NRAS, and probably PI3 kinase and Akt mutations, as well as PTEN and ligand expression levels to really crystallize the optimal patient that can benefit from EGFR antibody therapy. These are studies that are ongoing right now, and I think in the future will go beyond just using KRAS as a predictive marker for EGF receptor antibodies.

We're already using gene profiles at key cancer treatment centers, such as MD Anderson in Houston, Dana Farber in Boston, Memorial Sloan Kettering in New York, Vanderbilt in Nashville, and very soon at the Mayo Clinic. When patients with metastatic colon cancer come to these centers, they will have a profile of different mutations tested. PCR methodology is a cost-effective, robust, and reliable way to test for various mutations in one sample, and then hopefully be able to identify those patients who are most likely to benefit from EGFR antibodies.

For now, guidelines strongly recommend KRAS testing at the time of diagnosis in all patients with metastatic colorectal cancer.^[5,6] Under both sets of guidelines, anti-EGFR antibody therapy with cetuximab or panitumumab is recommended in the setting of wild-type KRAS disease.

Back to top

Answer: At this point in time, when we talk about biologic therapies for colorectal cancer we're mainly focusing on EGFR and the vascular endothelial growth factor (VEGF) bevacizumab. So one of the reasons why we must consider sequencing is that our current data show that we should not combine these agents. We have data from various trials that indicate that we should not combine panitumumab or cetuximab with bevacizumab in first-line therapy.

Now, the choice of therapy depends very much on the goal of therapy. In patients with advanced nonresectable colorectal cancer, our goal is palliative – therapy should extend the duration of life and maintain the quality of the life for as long as possible. In these patients, my goal is to make treatment as tolerable as possible with the emphasis on duration of therapy, because I strongly believe that patients benefit from a longer-term protracted therapy that may turn this disease from life-threatening disease to chronic.

So bevacizumab is a preferred first-line biologic in combination with a fluoropyrimidine like 5-FU or capecitabine.^5,7,8 This combination has been shown to prolong duration of response and PFS, and the incremental gains we see in PFS from line to line of therapy eventually leads to the improvement we've seen in OS over the past 10-15 years. For example, we started out with 12-month OS with just 5-FU. When we added irinotecan or oxaliplatin we went up to 15-18 months, and now we expect patients to live more than 2 years.

In terms of safety, we do have the concern of hypertension with bevacizumab. There is also the risk of stroke or other thromboembolic events, which are more common older patients (≥65 years of age).

For palliative care, I use cetuximab or panitumumab in later lines of therapy, some time after bevacizumab. In all the trials that we have, these EGFR antibodies retain their efficacy from line to line to line of therapy.⁵ Another consideration is the skin rash associated with EGFR antibody therapy. Patients don't want to be stigmatized early on when they start their therapy. To a lot of patients, it's important to continue to appear healthy, and not change the way they look. By reserving EGFR antibodies for later lines, many patients can go on for several years.

Now if we have a curative intent, and we need to rely on a strong anatomical response including measurable tumor shrinkage, we’ll want to use the EGFR antibodies in the earlier lines of therapy.^9-11 The addition of EGFR antibodies induces an enhanced response compared with chemotherapy alone. For example, in the CRYSTAL study, the response rate increased from 43% with FOLFIRI chemotherapy alone to 60% with the addition of cetuximab.¹⁰ In addition, EGFR antibodies might have an advantage in terms of downsizing tumor and rendering more patients resectable. This affects only about 10%-20% of patients overall, but in those patients, short-term treatment with an EGFR receptor antibody up front with the goal to induce the response and then send patients to surgery might actually be a better option than bevacizumab.⁵

This raises the issue of using EGFR antibodies in the neoadjuvant setting. Currently the role of these agents in the neoadjuvant setting is to downsize metastases so that patients who initially are unresectable might be candidates for surgery. Many of us still use chemotherapy up front even in resectable metastases to obtain biological information on these tumors, because we see that tumor shrinkage in a neoadjuvant situation is associated with better outcome. Results of the recent Cetuximab in Neoadjuvant Treatment of Non-Resectable Colorectal Liver Metastases (CELIM) trial, presented at the EMCC in Stockholm, confirmed that addition of cetuximab to standard chemotherapy was associated with longer overall survival in patients who had complete resection of colorectal cancer liver metastases and were initially deemed inoperable.^12,13 This study included 110 patients with unresectable metastatic disease who were randomized to receive cetuximab plus FOLFOX (leucovorin, fluorouracil, and oxaliplatin) or cetuximab plus FOLFIRI (leucovorin, fluorouracil, irinotecan). They were re-evaluated for resectability after 8 cycles and then every 4 cycles.

The use of either combination chemotherapy induced a 70% response rate in patients with KRAS wild-type tumors who were initially considered unresectable. As a result, 43% of these patients were resected, with 34% being R0 resections. (Table 2) Median overall survival in the overall study population was 33.1 months. In patients with KRAS wild-type tumors, median overall survival was 36.1 months. So those patients who did have resection had excellent outcomes, even if they recurred later. If we achieve a zero resection, essentially rendering patients free of disease, then patients actually benefit from longer overall survival compared to having no resection. When analyzed by R0 resection status, median overall survival was 46.7 months compared with 27.3 months in the group that was not R0-resected. The 4-year survival rate was 49% in R0-resected patients versus 16% in non-R0-resected patients (P = .002).

There was no significant difference between the chemotherapy regimens in terms of progression-free and overall survival rates.

Back to top

Answer: We have 2 new diagnostic approaches for these cancers: the use of positron emission tomography (PET) CT imaging, and a biomarker for trastuzumab. PET CT imaging is considered standard of care in most institutions. We perform a PET scan because we need to identify whether or not a patient has lymph node metastases or even peritoneal disease that is sometimes difficult to identify on a regular CT scan. We do routinely use PET CT imaging, as a diagnostic tool right before therapy in the gastroesophageal cancers.^14-18

Now there are studies even out there right now, based on some European data and some emerging American data that we can use PET CT imaging to identify response to therapy as early as 2 weeks after initiation of therapy. We can use this information to identify patients who might benefit from a change in the neoadjuvant strategy, which of course has curative intent. But this is something which is still being investigated in clinical trials.^19,20

I think what is more solid in terms of diagnostic approaches is testing for the HER2 biomarker, which allows us to identify which patients with gastroesophageal cancer and HER2 overexpression might benefit from therapy with trastuzumab. This is very much in line with what we've seen in breast cancer, where testing patients with breast cancer for HER2 overexpression and use of trastuzumab and other HER2 inhibitors is standard of care.

Our strongest data set comes from the international Trastuzumab in Combination with Chemotherapy Versus Chemotherapy Alone for Treatment of HER2-positive Advanced Gastric or Gastro-oesophageal Junction Cancer (ToGA) trial, which compared cisplatin plus a fluoropyrimidine with or without trastuzumab in HER2 overexpressing patients with gastroesophageal cancer, clearly put trastuzumab on the map in 584 patients with HER2-positive, inoperable, locally advanced, recurrent or metastatic gastroesophageal or gastric adenocarcinoma.²¹ These patients clearly had a benefit in overall and progression-free survival when trastuzumab was added to chemotherapy. (Table 3) The strongest correlation with outcome of the TOGA trial was seen in the patients with the strongest HER2 protein overexpression (immunohistochemistry [IHC] 3+ or IHC 2+ and fluorescence in-situ hybridization [FISH] positive). There was no difference in cardiac events between the 2 treatment arms.

That is what we are doing in the Mayo Clinic. We're selecting patients who have high protein expression, IHC 3+ or patients with IHC 2+ and FISH positivity. We do not use trastuzumab in patients who have no HER2 overexpression (IHC 0 or 1+ and FISH positive). We look for the higher overexpressors of HER2 and then consider them candidates for trastuzumab because those patients saw a clear benefit from adding trastuzumab.

Back to top