Digital Pathology and Biomarkers Help Pave the Way to Personalized Medicine

The biopharmaceutical industry is adapting to shrinking pipelines, increased market access barriers, significant financial strains and advances in technology. Companies must employ certain tools to bring life-saving medicines to market sooner and remain competitive in the rapidly changing arena of drug development.

Among the technological advances that can shorten drug development timelines and reduce the overall complexity of conducting clinical research, digital pathology and advanced biomarker development stand out. These areas hold the potential to not only reduce costs and timelines, but also deliver more targeted, and more effective therapies to patients.



Targeted Treatment
In 2008, researchers made a breakthrough when studying the V-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog (KRAS) gene - a molecular triggering gene that transforms a protein that is implicated in several types of cancers. They identified a mutation within this gene as a negative predictor for well-known oncology treatments, particularly with regard to a class of drugs called Epidermal Growth Factor Receptor (EGFR) inhibitors. This data proved to be compelling, as patients with the mutation have little chance of responding to this class of drugs. About 35 to 40 percent of colorectal cancer patients have these mutations. By excluding these patients, researchers can spare them any potential side effects and focus efforts on treating patients with wild-type KRAS.

Today, the KRAS mutation assay is used globally to select relevant patients for treatment with EGFR-targeted therapies. The American Society of Clinical Oncology and the National Comprehensive Cancer Network have recommended that all patients with metastatic colorectal cancer be tested for mutations to the KRAS gene.

The European Medicines Agency (EMEA) led this movement by officially recommending that only KRAS-wild-type patients be treated EGFR monoclonal antibody therapies. These recommendations further validate the role of mutation assays and biomarkers as cornerstones of personalized medicine and targeted therapies.

The KRAS breakthrough is an example of a molecular marker that is used to match the right patient with an approved targeted therapy, since virtually no patient who is harboring mutations in the KRAS gene responds to EGFR monoclonal antibody treatment. The targeted selection of patients is a role that biomarkers offer and will be explored by organizations that are developing targeted therapeutics in oncology.

Additionally, researchers are using biomarkers as a Pharmacodynamic (PD) tool to help to assess the efficacy of the therapeutic agent. This approach allows researchers and trial sponsors to understand how the drug is working and to do so earlier in the process. Multiple biomarkers are frequently interrogated to gain an understanding of the pathways that are being activated or inhibited in the cancer cell.

These panels of biomarkers are often a leading indicator of potential combination options either with chemotherapy or additional targeted agents. For instance, the Phosphoinositide 3-Kinase (PI3K) pathway is commonly invoked by tumor cells as a survival pathway.

By validating biomarkers that probe for loss of phosphatase and tensin homolog (PTEN) by Immunohistochemistry (IHC), assess amplification of the PI3K gene by Fluorescence In Situ Hybridization (FISH), or by evaluating the PI3K mutations by Reverse Transcription Polymerase Chain Reaction (RT PCR), developers of targeted agents against PI3K can assess early the efficacy of their PI3K inhibitor.

Then, when combined with biomarkers for another pathway, such as mammalian Target of Rapamycin (mTOR), for example, the accumulated data can provide insight into the value of combining PI3K and mTOR inhibitors for a particular tumor indication.



Digital Capture
Oncology biomarker work commonly involves tissue, specifically the stains on patient tissue that highlight the presence of a certain protein or genetic target. Digitally capturing this information allows an investigator to leverage the richness of data that is contained within the patient's tumor morphology.

Today's technology allows researchers to digitally capture an entire slide, share it almost in real time and then mine the image for quantitative and qualitative data. The ability to look at digital images from any investigative site worldwide, in essence, allows a researcher to use his or her computer as a microscope. It also eliminates issues that are associated with transporting and storing human tissue, and makes sample images available instantly, thereby providing cost savings and decreasing the overall development timeline.

The actual image analysis of tissue samples offers potential in the area of oncology even in its current embryonic stage. Fundamentally, image analysis can help to standardize the interpretation of IHC stains. As opposed to just looking at a picture and making an estimated guess as to how much protein is present, for example, digitized images enable pathologists to use sophisticated software to do this work in a quicker and more accurate manner.

Clinically, algorithms such as the Human Epidermal growth factor Receptor 2 (HER2), Estrogen Receptor (ER) and Progesterone Receptor (PR) have received regulatory approval to assist in patient diagnosis to help to identify those patients who might respond to a therapy, versus those who might not. These algorithms have the potential to limit the variability of interpretation among pathologists.

The use of digital image algorithms can also be extended to clinical trials. The ability to identify a specific patient population can help to reduce the overall time that is needed to conduct a trial. By knowing exactly what to look for on a slide sample, researchers can immediately narrow the potential pool of patients to a specific genotype or protein level.

Biopharma as a whole is beginning to embrace this emerging technology, and many organizations have established digital pathology networks with in-house scientists who are tuning digital algorithms to analyze samples.

It is conceivable that these activities may lead to an improved personalized medicine environment in which patients are selected more accurately to receive specific targeted therapies. Clinical pathologists will be able to furnish patient results to clinician colleagues with greater confidence, underpinned by algorithms that are designed to chase out confounding subjectivity.

Finally, a centralized digital pathology network in which all image analysis and readings are conducted in one location - using a standardized process - eliminates variability, increases accuracy, provides data storage security and enables the near real time dissemination of vital image data. By transferring highresolution images digitally, researchers will no longer need to ship physical specimens. This is particularly beneficial in countries that prohibit the export of human tissue, such as China.

As China continues to expand its involvement in clinical research, using digital pathology instead of traditional glass slides decreases development timelines by enabling pathologists anywhere in the world to analyze images and collaborate in real time. Working with an organization that has invested in a robust, scalable and secure network allows drug developers to reap these benefits.

Substantial thought has to be invested in such a network, but once it has been setup and deployed on a global basis, the drug developing organization can begin harvesting the benefits of expanded collaboration and reduced shipping and logistics costs. Best practices may also emerge. For example, digital pathology can enable the ability to use the best pathologist in the organization for a particular application (eg, a certain type of tumor or pathway).



Transforming Cancer Research
So, what does the biomarker and digital future look like? The evolution of biomarker and digital pathology technologies have the potential of enhancing the intelligent design of targeted oncology clinical trials while minimizing toxicity in patients. Meanwhile, the advent of personalized medicine pushes the need to develop more effective biomarkers, both for patient selection and for pharmacodynamic analysis. Therefore, the validation of these biomarkers becomes central to the success of deploying them in global clinical trials.

The utility of digital pathology and image analysis has the potential to assist in the successful use of these biomarkers in large, global Phase III trials. In the future, it may be possible to select the right patient with the optimal biomarker and to do so accurately with a helpful boost from digital image algorithms.

This can serve to increase the efficacy of the therapeutic agent and to screen out patients who are not likely to respond to a particular drug. At the same time, costs can be reduced by eliminating unnecessary clinical trials, shrinking development times, and avoiding the unnecessary treatment of patients who are not likely to respond.

In addition, by leveraging biomarker information, researchers can ascend to the next rung of patient treatment - understanding how to combine therapies. Through combinations - for example, chemotherapy plus a targeted therapy or two targeted therapies - researchers can begin treating patients by intelligently hindering the pathways that assist the tumor cell and activatingthe pathways that control it.

The patient will no longer simply be a "breast cancer" patient or "colon cancer" patient receiving routine prescriptions of the "typical" chemotherapeutic and targeted therapeutic cocktails. Instead, biomarker researchers will have the ability to decipher a pathway map of the patient and then rapidly identify the right combination of drugs to treat each individual patient - often in a blockade fashion to thwart the tumor cell's strategies - and to do so in a more cost-effective and time-effective manner.

Pharmaceutical companies recognize the potent power of biomarkers and have embraced the latter's utility in drug development. The recognition of digital pathology advantages is likewise developing rapidly. Regulatory agencies are strong advocates for biomarkers in drug development and have also initiated forums to highlight the potential uses of digital pathology and image algorithms for drug development.

These initiatives map a journey for cancer research and drug development, and the industry is committed to getting the best treatment to patients. Already, new chapters are appearing in the biomarker lexicon as investigators ponder on how to bring life and utility to emerging technologies. There is a desire to gain control of gene expression and to hunt down circulating tumor cells. And in line with the goals of personalized medicine, this is being carried out to achieve optimal patient outcomes, increased efficacy, decreased risk and reduced side effects for patients.