A new prognostic model has been developed to identify high-risk follicular lymphoma patients likely to progress early after receiving first-line therapy.
The nine-item Follicular Lymphoma Evaluation Index (FLEX) includes male sex, the sum of the products of lesion diameters (SPD) in the highest quartile, histologic grade 3A, more than two extranodal sites, ECOG performance status >1, hemoglobin <12 g/dL, β2 microglobulin >institutional upper limit of normal (ULN), peripheral blood absolute natural killer cell count <100/µL and serum lactate dehydrogenase >ULN.
Patients were categorised as low risk with scores 0-2 and high risk with scores 3-9.
The model was developed in a training cohort of 1,000 participants from the GALLIUM study and validated in a smaller cohort of 342 participants in the SABRINA trial.
The researchers, from Australia, UK, Switzerland, Germany and the US, found FLEX had greater accuracy for predicting progression of disease within 24 months (POD24) than existing clinical tools of Follicular Lymphoma International Prognostic Index (FLIPI and FLIPI-2) and the PRIMA-Prognostic Index (PRIMA-PI).
A high-risk score with FLEX was more sensitive in predicting POD24 than FLIPI and FLIPI-2 (60% v 53%) and more specific than FLIPI, FLIPI-2 and PRIMA-PI (68% v 59% and 47%) in the validation cohort.
“FLEX also demonstrated an improved ability to predict PFS compared with FLIPI-2 or PRIMA-PI in the validation (SABRINA) cohort, and a comparable ability to FLIPI,” the research team wrote in their paper published in the American Journal of Haematology.
It was also a robust indicator of PFS irrespective of first-line treatment regimen, they said.
“Importantly, FLEX is the first clinical score in FL to be developed in patients treated with bendamustine and either obinutuzumab or rituximab; it is therefore applicable to patients treated with the current standard of care.”
FLEX also had the advantage of easily measurable clinical components.
The study said it was not currently possible to create one integrated clinicogenomic model applicable to all patients as genes that may be prognostic in one treatment subgroup may not be prognostic in others.
“Moreover, the potential benefits of incorporating genomic data into a clinical model might be outweighed by increased cost and complexity.”
The study concluded that while its performance was still not optimal, FLEX was more accurate at discriminating which patients were likely to have poor PFS and OS than either FLIPI, FLIPI-2 or PRIMA-PI.
“It is important to be able to identify these high-risk patients at diagnosis, prior to treatment, as they may be candidates for alternative, risk-adapted therapies, which may include investigational regimens in clinical trials,” the study authors said.