JULY 9, 2019, New York- Bookies and pundits scour every evolving scrap of information to predict the outcome of the next game or election. These predictions can change on a dime, based on a player’s poor pass or a candidate’s stellar debate performance. Statisticians refer to the incorporation of continuously generated information as the calculation of in-game win probability.
Now a team led by Ludwig Stanford researchers Maximilian Diehn and Ash Alizadeh has taken a page from this playbook to generate more accurate prognoses for cancer patients. They’ve done so by designing a computer algorithm that can integrate many different types of predictive data-including a tumor’s response to treatment and the amount of cancer DNA circulating in a patient’s blood during therapy-to generate a single, dynamic risk assessment at any point in time during a patient’s course of treatment.
The new approach, named CIRI, for Continuous Individualized Risk Index, may also help doctors identify people who might benefit from early, more aggressive treatments and those who are likely to be cured by standard methods. The study was published online July 4 in Cell.
The researchers first looked at people previously diagnosed with diffuse large B-cell lymphoma (DLBCL). When a DLBCL patient is diagnosed, clinicians assess the initial symptoms, the cell type from which the cancer originated and the size and location of the tumor after the first imaging scan to generate an initial prognosis. More recently, clinicians have also been able to assess the amount of tumor DNA circulating in a patient’s blood after therapy begins to determine how the tumor is responding and estimate a patient’s overall risk of death. But each of these approaches assess risk based on a snapshot in time rather than a dynamic risk assessment that can be updated throughout the course of a patient’s treatment.
Alizadeh and his colleagues gathered data on more than 2,500 DLBCL patients from 11 previously published studies for whom the three most common predictors of prognosis were available. They used the data to train a computer algorithm to recognize patterns and combinations likely to affect whether a patient lived for at least 24 months after seemingly successful treatment without experiencing a relapse. They also included information from 132 patients for whom data about circulating tumor DNA levels were available prior to and after the first and second rounds of treatment.
CIRI’s prognostications were considerably better than that of existing methods. If a perfect score is assigned a value of 1, and the even odds of a coin-flip is 0.5, CIRI scored 0.8-which was not perfect but a considerable improvement on the current methods, which score about 0.6.
CIRI’s performance on data from previously published panels of people with a common leukemia, and another on breast cancer patients, was also encouraging. By serially integrating the predictive information over time, CIRI outperformed standard methods. Furthermore, the researchers found that it might be useful to identify patients likely to need more aggressive intervention within one or two rounds of treatment rather than waiting to see if the disease recurs.
More detail about these findings is available in the Stanford release from which this summary is derived.