Cancer Treatment and the use of Avatar Trials for Biomarkers...

Pharma Tech Outlook: Pharma Tech Magazine

Cancer Treatment and the use of Avatar Trials for Biomarkers Identification

Dr. Jean Pierre Wery, President, Crown Bioscience

Dr. Jean Pierre Wery, President, Crown Bioscience

The discovery of cancer biomarkers, novel cellular features specific to the tumor, has provided opportunities for understanding that cancer is not a single disease and that tumors deriving from the same tissue and presenting with similar histopathological features may not be the same. This has lead to improvements in the management of cancer patients by ensuring that the right sub-population of patients is treated with the most appropriate drug to fight their disease. Recent technological advancements have enabled the examination of several biomarkers at the same time and renewed interest in discovering new biomarkers. Biomarkers of cancer could include a broad range of biochemical entities as well as whole tumor cells found in circulation. A comprehensive understanding of the relevance of each biomarker is important not only for diagnosing the disease reliably, but also in helping to choose the best therapeutics that are more likely to benefit the patient.

“The discovery of new biomarkers of response and the accuratcy and reliability of existing biomarkers can be tested in Phase II-like avatar trials”

However sophisticated our biomarker discovery platforms have become; for some tumor types, the identification of ‘responders’ remains a challenge. Inaccuracy can cause suffering and unnecessary costs as some patients, thought to be responders, do not improve while those identified as ‘non-responders’ are not treated at all.

Recent studies have demonstrated the importance of avatar trials to validate existent or identify novel biomarkers of response to anti-cancer agents, before a single patient has been dosed. This has been shown to increase the speed at which techniques and treatments enter the clinic and reduce the high attrition rates currently associated with anti-cancer agents. This article will explore the latest studies on the use of cetuximab to treat Colorectal Cancer (CRC), the discovery of new biomarkers that predict response and the use of avatar trials to explore the accuracy of these biomarkers.

Avatar Trials

Cetuximab and Colorectal Cancer

CRC is one of the most common and deadliest malignancies with high frequency of metastasis (mCRC), affecting more than 1 million people globally every year. The common treatment options include a combination of different chemotherapy agents as well as more targeted immunotherapy approaches, such as the use of cetuximab or bevacizumab, monoclonal antibodies targeting the epidermal growth factor receptor (EGFR) .

Cetuximab was first approved for treating mCRC patients who show aberrant expression of EGFR. However, only about 10 percent of these patients would respond to cetuximab monotherapy as second line therapy. Reports have suggested that gene amplification and overexpression of EGFR and its ligands could potentially serve as the positive predictors of response, while other genetic alterations could serve as negative predictors. Nevertheless, with conflicting and inconclusive observations in the clinic, so far, it remains a challenge to effectively predict responses.

Activating mutations in a second oncogene, KRAS are currently used as exclusion criteria for cetuximab treatment since initial clinical results have shown that mCRC patients carrying these mutations are unlikely to respond to EGFR inhibition by cetuximab. Accordingly, genetic testing to confirm the absence of KRAS mutations (and so the presence of the KRAS wild-type gene) is now clinically routine before the start of treatment with EGFR inhibitors. Recent data suggest that around 65 percent of mCRC patients have a wild-type KRAS gene, and therefore are expected to respond . mCRC patients with wild-type KRAS tumors have shown in trials an overall response rate of over 60 percent and a decreased risk for progression of over 40 percent when treated with cetuximab.

However, recent studies seem to contradict these data and indicate that the current cetuximab label regarding KRAS mutation may be incorrect, or at least inaccurate. This study unexpectedly found that some patients with a mutated form of KRAS responded to treatment, while only 35-50 percent of CRC patients with wild-type KRAS responded. Therefore, it is of medical importance and urgency to find novel biomarkers that are better predictors of response than KRAS mutational status to be able to include patients erroneously considered ‘non-responders’ and avoid unnecessary costs and suffering.

PDX models

The discovery of new biomarkers of response and the accuracy and reliability of existing biomarkers can be tested in Phase II-like avatar trials. In these trials, Patient-Derived Xenografts (PDX) are used as patient avatars. PDX models are generated during human tumor specimens that have never been passaged in plastic. This helps preserving important tumor characteristics such as the original tissue architecture, the presence of vasculature, the heterogeneous nature of the tumor, and all its molecular and genetic alterations. These models, used in human surrogate trials, enable the identification of the right compound for the appropriate patient population, before new drugs are tested in the clinic, with a higher correlation with clinical outcomes compared to conventional cell line-derived models.

In a study performed by Crown Bioscience, a panel of CRC-PDXs was established to evaluate their response to cetuximab and investigated the reliability of KRAS mutation as a biomarker predictive of response. CRC is among the cancer types that are most readily engrafted into immuno-compromised mice with high take-rate. The results from our study stratified the CRC-PDXs between responders (30 percent) and non-responders (70 percent) to cetuximab. Because of this distinct pattern of responses it was possible to further investigate the presence of KRAS mutations in these models and correlate it with sensitivity to treatment. Our study suggests that since KRAS activating mutations are found in both responders and non-responders, the role of KRAS as biomarker of response to cetuximab in CRC patients should be revaluated.

Better Biomarkers

Although KRAS mutations failed to efficiently predict response to cetuximab treatment in avatar trials, thanks to the genetic profiling of the PDX models enrolled in the study it was possible to discover a new set of oncogenic mutations whose presence showed a stronger correlation to cetuximab resistance than KRAS mutations. Work is currently in progress to confirm this correlation on an independent cohort of CRC-PDX.


PDX are experimental models that can closely mimic patient tumors. The use of avatar trials can increase the speed at which drugs enter the clinic and reduce attrition rates, enabling more efficient treatment of patients. Through the use of avatar trials, Crown demonstrated how biomarkers once thought to predict negative response, may in fact not be as accurate, meaning that patients who were excluded from treatment could have at least in part benefitted.

Weekly Brief

Read Also

Managing a method transfer:  How to achieve a smooth transition to manufacturing

Managing a method transfer: How to achieve a smooth transition to manufacturing

Cory Nelson, Director of Contract Manufacturing, WuXiAppTec
Analytics as a competitive advantage takes 3 things - access, cleanliness, and action

Analytics as a competitive advantage takes 3 things - access, cleanliness, and action

Matt Southwick, Executive Director Clinical Systems and Analytical Reporting at Amgen
Tackling Ai's Control Problem

Tackling Ai's Control Problem

Dr Harvey Lewis, Associate Partner and Chief Data Scientist, UK & Ireland, EY
The Modern Day Biomechanics

The Modern Day Biomechanics

Julio Martinez, Medical Director, Baystate Health
Clinical Trial Analytics Prescription For Bio-Pharmaceutical R&D Industry

Clinical Trial Analytics Prescription For Bio-Pharmaceutical R&D Industry

Ankit Lodha, Associate Director, Clinical Analytics & Innovation, Takeda