In recent years, great strides have been made in understanding and treating lung cancer. While we’re excited to see these many developments, our priority at LUNGevity is to ensure patient voices, priorities, and experiences are considered as part of these improvements. In the clinical trial space, despite more calls for greater patient centricity, there are still areas lacking, such as how patient-reported data are analyzed and presented.
Cancer clinical trials today tend to focus on overall survival (length of time from start of treatment that patients are still alive) and progression free survival (length of time during and after treatment that a patient lives with the disease but it does not get worse) as primary endpoints. However, progression free survival hasn’t always translated into living longer. Both patients and doctors want to know whether the quality of life of the trial participants was affected by the new treatment.
The problem is that quality of life can be difficult to interpret; what one person defines as important to them will be different to the next person. But that is ok; there are more concrete, yet related, concepts. For example, we can measure how a treatment affects someone’s ability to do normal daily activities like walking (also known as physical function in clinical trial terms) and whether it improves, stays more or less the same, or gets worse. The Food and Drug Administration (FDA) Oncology Center of Excellence is particularly interested in how physical function changes during treatment for patients who are enrolled in cancer clinical trials.
On July 23rd, I presented some of my work at the FDA’s Clinical Outcome Assessments in Cancer Clinical Trials workshop. I took what I see as being an oversimplification in how clinical trial survey scores for concepts like physical function, created by tallying up patients’ responses, are often presented to regulators and other scientists in literature. I have an example of what is often shown to below.
This chart compares the average physical function score (mean) for patients in the treatment arm (investigational therapy) to those in the active control (standard of care) over time on trial. The lines sticking out of the dot tell us the confidence interval. The longer it is, the less confidence we have in the average; the smaller it is, the more confidence we can have. However, there is no average patient, but we also can’t tell regulators, payers, and the scientific community every patient’s unique story. So how do we find a compromise?
The solution I presented was to look at the scores broken into 4 equal quarters and track how scores changed over the first 6 months of treatment. These are called Sankey diagrams and were initially developed by Captain Matthew Sankey to depict energy efficiency of a steam engine; it can be good to pull ideas from other disciplines. The same data is presented below in one of these Sankey diagrams. Of course, other scientists can choose different groupings for scores that are meaningful to their clinical trial as the principle still holds.
This graph provides a lot more information than the one that shows the average scores but uses the same data. And while this might be complicated to look at, with each color and curve representing something different, that’s the point: the graphic is so much richer in information.
In this graph we can see that some trial participants stay more or less the same for their physical function, some get worse, some come off treatment, and a small group do actually better on physical function. We can conclude that while around half (56%) of the participants started in the top quarter of the physical function score, this rapidly dropped off over the first 6 months of treatment; many patients, regardless of where they started, experienced decline in physical function. However, there were a small number of patients reporting better physical functioning. From the first chart, all we can say is that on average patients physical functioning showed a slight decline while they received treatment.
Like all analyses, there are strengths and weaknesses, and this is no exception. The biggest strength is that all information regarding physical function can be presented, which is especially helpful to researchers and regulators in assessing the quality of the data (i.e., was there much missing). The biggest weakness, though, is that there is a lot of information here to process.
Why does this matter? Well, pictures tell us a story. In the first picture, there isn’t much detail, and we all know a good story requires some details. The second picture might take longer to take in the story, but we can see the proportions of people whose physical functioning changed; we also know the proportion of people whose treatment ended. A physician could use the second picture to let a patient know that most patients on the trial experienced a decline in their ability to do activities like take a long walk. But that wasn’t true for all patients: some stayed about the same and some even did a little better. This can help improve shared decision-making.
To help get researchers, regulators, and clinicians accustomed to Sankey diagrams, we will be using these in our presentations of data collected as part of Patient-Focused Research Center initiatives at meetings we attend. I believe that once people get comfortable looking at these pictures, they will see the value they have to offer.
- ASCO 2021: Progress in Lung Cancer Treatment
- A Lung Cancer Patient’s Experience with Palliative Care
- Coming to Terms with a Lung Cancer Diagnosis
Dr. King-Kallimanis manages LUNGevity’s Patient-focused Research Center (Patient FoRCe), which aims to connect the patient voice with healthcare professionals, regulators, policymakers, and developers of drugs to ensure that their voices are heard and incorporated into decisions. Dr. King-Kallimanis’s work improving clinical trial physical function data presentation helps ensure patients can make better informed decisions about the treatment right for them based on their individual preferences.