Cancer therapies that harness a patient’s own immune system to improve the body’s ability to fight off cancer have shown promising outcomes. One such immunotherapy is chimeric antigen receptor (CAR) T-cell therapy which takes a patient’s own immune cells, T cells specifically, and engineers them to find and kill the cancer cells. CAR T-cell therapy has particularly shown success with treating blood cancers. While this novel treatment has shown a lot of promise, it’s not without potential side effects. Immune effector cell-associated hematotoxicity (ICAHT) is one possible adverse effect of CAR T-cell therapy. ICAHT results in patients developing “low blood counts, which can last for weeks to months following CAR T-cell therapy,” states Dr. Emily Liang, a hematology/oncology fellow at the University of Washington and Fred Hutchinson Cancer Center mentored by Fred Hutch Associate Professor Dr. Jordan Gauthier. She continues, “this can lead patients to be dependent on transfusions and increases their risk of infection. Infection is the most common cause of non-relapse death after CAR T-cell therapy, so a better understanding of how ICAHT develops and how to treat it is very important.”
To classify the severity of ICAHT, the European Hematology Association and European Society of Blood and Marrow Transplantation created an ICAHT grading system which is “based on the degree and duration of neutropenia (low absolute neutrophil count [ANC]) following CAR T-cell therapy,” explains Dr. Liang. Neutrophils are immune cells critical for fighting off infections. When a person has too few neutrophils, they are at a higher risk of getting an infection, and for cancer patients, getting an infection can be life-threatening. Low ANC values are also the defining feature of ICAHT. “Depending on how low the ANC is and how long the ANC stays low, patients are assigned an ICAHT grade. As you may imagine, manually checking the laboratory values and counting the number of days of low ANCs for each patient, then doing that for the hundreds of patients who may be in a research study, is very time-consuming and prone to error. Thus, we wanted to create an automated way to grade ICAHT,” shares Dr. Liang. Alongside researchers from the Gauthier group, Dr. Liang helped develop an automated ICAHT grading method, which was recently published in Bone Marrow Transplantation.
To develop an automated grading system, the researchers adapted existing computational methods that are used for other disciplines. The central part of their automated grading method relies on heatwaveR, “a climate research package in [the coding language] R that allows the user to compute the duration that the temperature of the ocean exceeds or is under a certain threshold. We adapted this method to compute the duration that a patient’s neutrophil count is below the ICAHT grading thresholds,” states Dr. Liang. With this automated program, a healthcare provider can use an Excel spreadsheet that includes the dates of when blood work was collected within the first 30 days post-CAR T-cell infusion and the ANC values that were measured from those blood draws as the input information. Since ANC values are important in assessing ICAHT, bloodwork measuring ANC would have been collected frequently during the initial 30-day period, but likely not every day. Using this spreadsheet with a patient’s ANC values that were measured during the 30-day post-infusion period as the input for this automated computation method, the program imputes, or estimates, any missing ANC values from days where blood work was not collected so that values exist for every day within the 30-day time frame.
In developing this method, Fred Hutch researchers got to work with their “colleagues at Memorial Sloan Kettering Cancer Center, including Dr. Kai Rejeski, who helped create the ICAHT grading system. We enjoyed sharing feedback about our method to make it as useful and accurate as possible. We will be working more together on research related to ICAHT!” After generating this automated system, the research team tested it out on data from over 1,200 patients from two independent cohorts being treated at Fred Hutchinson Cancer Center or Memorial Sloan Kettering Cancer Center. The ICAHT grading for these patients was performed both manually and computationally, which demonstrated high concordance with manual grading.
Dr. Liang emphasizes that the method’s biggest strengths “are that it is automated, fast, and very accurate. Compared to manual grading, which can take many hours, our method can be easily applied in the R programming language and can assign ICAHT grades to hundreds of patients in seconds. People who are not familiar with coding in R may be intimidated, but our paper provides a step-by-step tutorial for implementing our method so that it can be used by anyone!” Furthermore, she adds, “having a reliable ICAHT grading system means that providers and researchers can more easily and consistently assign ICAHT grades. This will standardize and encourage the reporting of ICAHT in both retrospective studies and clinical trials, which will further improve our understanding of this serious side effect.” Future research from Dr. Liang and the Gauthier group will use this method to better predict who will develop severe ICAHT. Dr. Liang notes that they can also adapt their “method to other treatments, such as hematotoxicity after stem cell transplantation.”
This work was supported by the National Institutes of Health, Swim Across America, the School of Oncology of the German Cancer Consortium, the Munich Clinician Scientist Program, the Bruno and Helene Jöster Foundation, and the Bavarian Center for Cancer Research.
Fred Hutch/University of Washington/Seattle Children’s Cancer Consortium members Drs. Jordan Gauthier and Andrew Portuguese contributed to this study.
Liang EC, Rejeski K, Fei T, Albittar A, Huang JJ, Portuguese AJ, Wu Q, Raj S, Subklewe M, Shouval R, Gauthier J. 2024. Development and validation of an automated computational approach to grade immune effector cell-associated hematotoxicity. Bone Marrow Transplant.