February 2013, Vol 2, No 1

← Back to Issue

Progress in Personalized Approaches to Colorectal Cancer

Alice Goodman


Advances in the understanding of the molecular basis of colorectal cancer were featured at a pre-meeting Press Cast for the 10th Annual GI Cancers Symposium held January 24-26, 2013. Three separate presentations focused on: a new molecular classification system for colorectal cancer; gene expression profiling of circulating tumor cells as a prognostic strategy in pancreatic cancer; and the benefit of surgery after treatment with imatinib in gastrointestinal stromal tumors (GIST).

New Molecular Subtypes of Colorectal Cancer

Molecular profiling has identified subtypes of breast cancer that have treatment implications. A study presented here showed that molecular profiling can also be applied to colorectal cancer, identifying 3 distinct gene expression patterns that are associated with different prognoses and responses to adjuvant chemotherapy.

“This study clearly shows that there are different subtypes in colorectal cancer with completely different biological and clinical characteristics. We hope that with continued research, we’ll be able to develop new molecular tests based on this classification system, not only to identify patients needing more aggressive treatment, but also to predict which patients will respond to specific chemotherapy drugs and targeted agents, regardless of cancer stage,” stated study coauthor Josep Tabernero, MD, director of clinical research at Vall d’Hebron Institute of Oncology in Barcelona, Spain.

Defining these subtypes may be especially helpful for stage II patients, since it is difficult to predict risk of relapse in this group of patients using clinical and pathological criteria. Gene array tests such as Oncotype DX and ColoPrint can be helpful, but it is not clear which patients should receive additional treatment, and these tests do not provide actionable information after relapse.

At present, there is no good way to select treatments for individual patients, Tabernero explained. KRAS status is the only predictor of anti-EGFR treatment activity.
Gene expression data were collected from 188 colo­rectal cancer patients stages I through IV to develop a classification system based on hierarchical clustering of genes. Three types of gene signatures were found to be specific for subtype A, subtype B, and subtype C colo­rectal cancer, and the subtypes were found to correlate with benefits from 5-fluorouracil (5-FU)-based therapy.

The subtypes were characterized as deficient epithelial subtype A, associated with a good prognosis (no adjuvant treatment or 5-FU–based treatment needed); proliferative epithelial subtype B; and mesenchymal subtype C. Both subtypes B and C are associated with poor prognosis; subtype B is chemotherapy responsive, while subtype C is chemotherapy resistant, suggesting that new targeted therapies are needed.

The system was then validated in tumor samples from 543 patients with stages II and III colorectal cancer. The breakdown of tumor subtypes was: subtype A in 21.5%, subtype B in 62%, and subtype C in 16.5%.

Ten-year follow-up showed that subtype C had the worst outcome, with no benefit from chemotherapy. Subtype A had a good prognosis with or without adjuvant chemotherapy. Subtype B had a significantly improved outcome with chemotherapy versus no

These subtypes are currently being evaluated in stage IV colorectal cancer, he said. “I predict that within the next 2 or 3 years, we and other research teams will have several gene expression signatures developed,” Tabernero said.

Gene Expression Profiling of CTCs in Pancreatic Cancer

Preliminary results from an ongoing prospective study suggest that gene expression profiling of circulating tumor cells (CTCs) in the bloodstream of pancreatic cancer patients may enable selection of personalized chemotherapy. The study suggests that specific gene expression profiles detected in CTCs may predict response to chemotherapy regimens. The researchers also identified key molecular pathways implicated in the development of treatment resistance and worsening cancer.

Survival remains poor in pancreatic cancer, with less than 5% of patients alive 5 years after diagnosis. New chemotherapy treatments are emerging, but as yet there are no validated biomarkers for treatment selection. CTCs are already being applied in breast cancer, and they would be useful as biomarkers for pancreatic cancer because it is difficult to biopsy. The study, reported by Kenneth Yu, MD, Memorial Sloan-Kettering Cancer Center, New York City, is the first to use pharmacogenomics modeling to analyze genetic response to chemotherapy regimens.

“Pharmacogenomics modeling is a promising and exploratory tool for predicting treatment response in pancreatic cancer. This technique can be performed at initial diagnosis and at disease progression. Gene pathway analysis provides insights into pancreatic cancer prognosis and the development of drug resistance,” Wu said.

The study enrolled 50 patients with unresectable stage II-IV pancreatic cancer. Patients were treated with 12 different chemotherapy regimens according to investigator’s choice. CTCs were collected prior to chemotherapy and at disease progression with Vita-Cap. Total RNA was extracted, and gene expression profiling was performed. A pharmacogenomics model for the 12 chemotherapy regimens was applied to gene expression profiles to predict sensitivity to chemotherapy. Investigators compared the gene expression differences in patients who continued to respond to chemotherapy and those who progressed on treatment.

“The questions we asked were: 1) does the pharmacogenomic model predict treatment response and resistance, and 2) does the pharmacogenomic profile change when cancer progresses,” Wu explained.

Among 20 patients who experienced disease progression, 6 had been treated with chemotherapy that the model predicted they would be sensitive to; 6 received treatment predicted to achieve an intermediate response; and 8 received treatment that was predicted to be resistant.

“Patients treated predicted by our model to be more effective did better,” Wu said. Median time to progression, according to a preliminary analysis, was 7.3 months in the sensitive group, 5.3 months in the intermediate group, and 3.7 months in the resistant group.

The chemosensitivity patterns were changed at disease progression, Wu continued, reflecting the development of resistance.

A further analysis showed that the E2F1 pathway disruption was associated with longer treatment response, while disruption of the NF-?B pathway was associated with a shorter treatment response. Both the phospholipase C and retinoblastoma 1 pathways were disrupted at disease progression.

The ultimate goal is to select the best treatment strategy for the individual patient, and to monitor patients during the course of therapy so that treatment can be adjusted according to molecular signs of worsening, he said.

Surgery After Imatinib

Imatinib was one of the first major advances in targeted therapy, providing the first life-extending
effective treatment for GIST and chronic myeloid leukemia. A retrospective study presented here found that patients whose residual lesions were resected after treatment with imatinib had improved survival compared with those who did not undergo resection.

“This treatment strategy is worth trying as a clinical practice at centers with experienced surgical teams,” stated Seong Joon Park, MD, Asan Medical Center, Seoul, Korea. He noted that the study was retrospective and therefore cannot establish a treatment guideline, but it is supportive evidence for a strategy that is already being used.

The study enrolled 134 patients with metastatic or recurrent GIST treated with imatinib alone or imatinib plus surgery to resect residual tumor, followed by additional imatinib. Patients were eligible for the study if they had maintained a response to prior imatinib for at least 6 months.

Surgery was performed in 42 patients, and 92 patients were treated with imatinib until disease progression.

Patient and tumor characteristics were similar in the groups, with 2 exceptions: median age was 51 years in the surgically treated group and 58 years in those treated with imatinib alone; and peritoneal metastases were found in 12 and 56 patients, respectively.

Median progression-free survival was 87.7 months in the patients who had surgical resection of residual disease after imatinib versus 42.8 months in those who were treated with imatinib alone (P=.001). Median overall survival had not yet been reached in the surgical group and was 88.8 months for imatinib alone (P=.001).

Park said that about one-third of patients with residual disease can have a successful resection. “It is often recommended in younger patients with a good performance status. Surgery can be difficult if multiple peritoneal metastases are present,” he noted.

Multivariate analysis showed that low initial tumor burden was associated with longer survival, while low initial tumor burden, female gender, and KIT exon 11 mutation were associated with delayed disease progression. The KIT mutation is also associated with response to therapy.

Uncategorized - February 20, 2013

Applying Personalized Medicine in the Overall Healthcare Environment

Dear Reader, Welcome to the second year of publication of Personalized Medicine in Oncology (PMO)! We launched this journal with the goal of providing the practicing oncologist with a translational guide to personalized medicine’s research advances and system improvements that bring targeted treatment into practice, thereby transforming personalized medicine from [ Read More ]

Uncategorized - February 20, 2013

Women With Triple-Negative Breast Cancer Highly Likely to Carry BRCA1 Gene

Triple-negative breast cancer (TNBC) was strongly associated with BRCA1 status, but not with BRCA2 status, in a large study of medically insured women. The study showed that the number of patients with BRCA mutations with a TNBC profile is statistically significant. The authors of this poster, presented at the 2012 [ Read More ]