Risk management in cancer clinical trials – new recommendations for imaging biomarkers

In recent years, there has been rapid growth in the use of molecular and functional imaging for clinical cancer research. Experts have focused their attention on devising risk management methodologies to optimize the use of such modalities. However, the rapid pace of progress in technology and new insights into cancer mean that risk management itself is permanently under re-development.

RECIST: driven by growth of imaging techniques
At present, the most widely used tool to assess treatment response and risk versus benefit in clinical trials is known as RECIST (Response Evaluation Criteria In Solid Tumors).
The original RECIST criteria date to the year 2000, following a collaborative effort between the US National Cancer Institute (NCI), the European Organization for Research and Treatment of Cancer (EORTC) and the Cancer Institute at Canada Clinical Trials Group.
RECIST was the first major effort since World Health Organization definitions in 1979, published in the WHO Handbook. By the early 1990s, the WHO definitions began to confront major limitations, above all with respect to changes in the size and minimal number of ‘measurable’ lesions. In turn, the key reason for this was the rapid development and growth of new imaging techniques such as CT and MRI, as well as confusion about how to integrate their far more precise data into response assessments.

Criteria are tumour-centric, not patient-centric
RECIST criteria define when treatment leads to tumours improving, stabilizing or progressing. They are principally intended to evaluate tumour response as a prospective endpoint in clinical trials. Although concurrent benefits are obtained by clinicians who use imaging studies to determine the success of a particular therapy and whether to continue or discontinue it, this is not the intention of RECIST. Indeed, one of the most significant elements of RECIST is that the criteria are tumour-centric, rather than patient-centric.

NCI guidelines in mid-2000s
The mid-2000s witnessed some technology-specific imaging guidelines for clinical trials. The best known of these followed workshops by the Cancer Imaging Program of the US National Cancer Institute.
They included guidelines on dynamic contrast MRI (DCE-MRI) and on the use of 18F Fludeoxyglucose (FDG) PET as an indicator of therapeutic response in clinical trial patients.

2009: RECIST 1.0 upgraded to 1.1
The RECIST criteria were updated in 2009. While the original set are now commonly known as RECIST 1.0, the revised criteria are called RECIST 1.1.
The reasons for the revisions were manifold. During the 2000s, numerous prospective analyses confirmed the validity of substituting uni-dimensional for bi-dimensional or even three-dimensional criteria. In spite of some exceptions (such as mesothelioma), uni-dimensional criteria seemed to perform adequately in solid tumour studies up to phase II.
However, several questions had also arisen over the past decade and these demanded further clarity. Key questions, which the revised RECIST guidelines address, include:

• whether it was possible to assess fewer than ten lesions without affecting the overall assigned response for patients
• making an assessment of lymph nodes
• ways to apply RECIST in randomized phase III trials where progression rather than response is the primary endpoint (especially for patients who do not demonstrate measurable disease)
• the applicability of RECIST in trials of targeted non-cytotoxic drugs
• the need for response confirmation

Qualified endorsement of FDG-PET
Finally, RECIST 1.1 provides guidelines on newer imaging technologies, especially FDG-PET (where preliminary efforts on its use as a “qualified biomarker” had already been made by the Cancer Imaging Program of the US National Cancer Institute in 2006).
Although RECIST 1.1 cautions that FDG-PET response assessments need additional study, the revised criteria do note that “it is sometimes reasonable to incorporate the use of FDG-PET scanning to complement CT scanning in assessment of progression (particularly possible ‘new’ disease). RECIST 1.1 also provides an algorithm to identify new lesions from FDG-PET imaging.
As mentioned previously, one of the key questions considered by the Working Group revising RECIST was to determine whether it was appropriate “to move from anatomic unidimensional assessment of tumour burden to either volumetric anatomical assessment or to functional assessment with PET or MRI.” The Working Group concluded that there was (still) not enough standardization or evidence “to abandon anatomical assessment of tumour burden.” The only exception was to use FDG-PET imaging as an adjunct to determination of progression, and pursue appropriate clinical validation studies for new technologies.

Imaging hybrids such as PET-CT, however, remain strongly qualified. The revised criteria state that “the low dose or attenuation correction CT portion of a combined PET–CT is not always of optimal diagnostic CT quality for use with RECIST measurements.”

Immunotherapy poses specific challenges
Almost in parallel to the RECIST revision, another recent initiative has sought to take into account the fast-emerging new field of immunotherapy.
Immunotherapy is also known as biological therapy and biotherapy and uses the body’s immune system to produce anti-tumour effects and fight cancer. It works by either stimulating a patient’s immune system to attack cancer cells or providing the immune system with what it needs, such as antibodies, to fight cancer. 
Examples of  immunotherapeutic agents include monoclonal agents, cancer vaccines and man-made versions of cytokines, the chemical agent in immune cells.
However, the clinical response to an immunotherapeutic agent can sometimes manifest only after “an initial increase in tumour burden or the appearance of new lesions (progressive disease).” In other words, such drugs would fail in clinical trials which measured response using WHO or RECIST criteria, because they fail to take account of the time gap in many patients between initial treatment and the apparent action of the immune system to reduce the tumour burden. Such a failure would occur in spite of the fact that these drugs ultimately prolonged life.

irRC Criteria
The so-called irRC (Immune-Related Response Criteria) consists of a set of published rules for evaluating anti-tumour responses with immunotherapeutic agents. It seeks to define when tumours respond, stabilize or progress during treatment. irRC was developed by a team of researchers from the US and Austria, France, Germany and Italy on the basis of experience with the CTLA-4 function blocking antibody ipilimumab in phase II trials in patients with advanced melanoma. Other immunotherapeutic anti-cancer drugs to have recently been approved in the US and Europe include pembrolizumab, sipuleucel-T and nivolumab. These drugs offer considerable promise for patients with advanced lung cancer, prostate cancer, renal cell carcinoma and melanoma.

Molecular and functional imaging
The advent of new drugs offering new promise for some of the toughest cancers is likely to continue in the coming years. Meanwhile, imaging biomarkers are expected to provide invaluable information on disease staging and characterization. Increasingly, researchers are enthused about the fact that molecular and functional imaging permits the detection or absence of response within days after the onset of treatment. In turn, early stage detection makes it feasible for non-responding patients to avoid unnecessary toxicity associated with therapy.
The development and and use of imaging biomarkers, however, is a complex process with complex data acquisition methods as well as specific regulatory issues for the trialling of new imaging agents. Quality control issues associated with image processing and the use of multi-vendor software are major, additional challenges – which can impact on data standardization. Such problems are clearly going to be more acute in the case of multicentre international trials.

Improving the use of imaging biomarkers
At the end of 2015, experts from Europe and the US published a paper in ‘Lancet Oncology’, aimed at “improving the implementation and utilization of imaging biomarkers in cancer clinical trials.” The authors represented the EORTC and the United States NCI – both of which have been associated with RECIST – as well as the European Society of Radiology (ESR) and the European Association of Nuclear Medicine (EANM).

The authors explicitly noted the promise of novel imaging modalities on disease staging and characterization and the rapid detection interval offered by molecular and functional imaging. Their aim is to “propose a practical risk-based framework and recommendations on imaging biomarker-driven trials that allows identification of risks at trial initiation so that resources can be better allocated and key tasks prioritized.” The paper also recognizes the “essential roles” played in clinical trials by other stakeholders such as “regulatory bodies, pharmaceutical companies, and patients.”
Dr Yan Liu, lead author of the paper and the Head of Translational Research, Radiotherapy and Imaging at EORTC, noted that cancer clinical trials have always sought to find a right balance between maximizing data quality and minimizing cost. Here, he said, “risk management can be an extremely helpful tool, because it can help us to prioritize, reduce costs, and decrease attrition rates.”

The Risk Assessment Plan would be best realized via a multi-disciplinary team which includes imaging experts, oncologists, as well as study project managers.
It should be reviewed and updated throughout a trial.

Towards personalized medicine
The need for robust risk management approaches in imaging was illustrated shortly before publication of the above paper in ‘Lancet Oncology’. On December 9, scientists at the University of Manchester and the Institute of Cancer Research in London announced development of a new oxygen-enhanced MRI test which mapped areas of hypoxia (or oxygen deprivation) within tumours – often a sign that a cancer is growing aggressively. The aim of the test is to enable doctors identify more dangerous tumours before they spread around the body – and tailor treatment accordingly.
Researchers used the technology to produce hypoxia maps within tumours in mice. It is now being further developed through clinical studies of cancer patients. According to study co-author, Dr. James O’Connor of the University of Manchester: “There is currently no validated, affordable and widely available clinical imaging technique that can rapidly assess the distribution of tumour hypoxia.” He hoped that oxygen-enhanced MRI will not only help identify the most dangerous tumours, but also assist in the monitoring of treatment response.
On his part, Nell Barrie of Cancer Research UK noted that “this early-stage research in mice will help to find new ways to use existing scanning technology to monitor and personalize each patient’s treatment …”.