Primary Pulmonary Diffuse Large B-Cell Lymphoma: A Predictive Model
—This large, population-based study of patients with PP-DLBCL assessed prognostic factors and confirmed that certain common therapeutic interventions can modify disease progression.
Being age 65 years or older, male, and having liver metastases increases the likelihood of a poor outcome in patients with primary pulmonary diffuse large B-cell lymphoma (PP-DLBCL), according to new research on prognostic factors for the disease.1 The study—the first large, population-based trial of its kind in PP-DLBCL—also confirmed that commonly used interventions, such as radiotherapy, surgery, and chemotherapy, can modify disease progression, leading the investigators to develop a predictive model that offers good prognostic discrimination at 1 and 3 years.
In approximately one-third of patients with DLBCL, the most common form of non-Hodgkin lymphoma, the disease originates in extranodal sites. In a previous Chinese study, the most common such sites were the stomach (22.4%), intestines (16.0%), noses and sinuses (8.9%), testes (8.4%), and skin (7.9%). PP-DLBCL accounted for just 1.5% of such cases.2 A study of data from US patients found that, when compared with other sites of origin, extranodal DLBCL in the gastrointestinal tract, liver, and pancreas had a worse prognosis. Disease originating in the head and neck, skin, soft tissues, and bone had a better prognosis.3
.jpg)
The 5-year progression-free survival (PFS) is approximately 33% for PP-DLBCL, which is lower than the 49% overall 5-year PFS rates associated with extranodal DLBCL.
As a rare entity, PP-DLBCL is less well studied than the more common extranodal forms of DLBCL. There are no widely recognized predictive models to assess PP-DLBCL prognosis. The goal of the new study was to employ data from the Surveillance, Epidemiology, and End Results (SEER) database to construct such a model.
For the analysis, 825 patients diagnosed with PP-DLBCL between 2010 and 2019 who were represented in SEER were identified and their data were analyzed. Variables and outcomes collected include age at diagnosis, clinical and pathological information, survival time, primary site, intrapulmonary and extrapulmonary metastases, and therapies received.
On univariate and multivariate analysis of prognostic factors in a training group analysis with 577 of the patients from the SEER database, older age and extrapulmonary metastases were identified as adverse prognostic factors. Surgical intervention, chemotherapy, and radiotherapy were found to be favorable prognostic factors.
All of the prognostic factors were entered into a nomogram to predict overall survival (OS) at 12, 24, and 36 months and tested in a validation group of 248 patients.
Area under the curve (AOC) calculations for the prognostic factors for OS at 12, 24, and 36 months in the training group were 0.765, 0.762, and 0.751, respectively. In the validation group, the AOCs were 0.774, 0.762, and 0.756, respectively. The good predictive accuracy and high agreement between predicted and actual observations suggest that the model developed by the researchers will perform well for predicting outcomes in patients with PP-DLBCL.
Absent a predictive model scoring system specific to DLBCL, the International Prognostic Index (IPI) is an option,4 but the authors of this study, led by Zhiping Deng, MD, and Maoliang Tian, MD, both of the Zigong First People’s Hospital, Zigong City, China, asserted that their model is more accurate and applicable.
Pointing to the inclusion of organ metastases and treatment as variables in the prognosis, Dr. Tian reported that “our model includes factors not found in the IPI score and is able to more comprehensively assess the prognosis.”
Published:
References
