Novel Simple Prognostic Index (NJHSG-PI) Can Effectively Predict Prognosis in Patients with Newly Diagnosed DLBCL : A Real World Analysis across 14 Cancer Centers in Japan

[Introduction/Purpose]
Maurer et al (J Clin Oncol, 2014) have shown an overall survival of American and French patients with diffuse large B cell lymphoma (DLBCL) who achieved event-free survival at 24 months (EFS24) to be equivalent to that of the age- and sex-matched general population (GP). But it has not been rigorously validated in other countries yet and it has not been reproducible in some reports. Therefore, we compared the overall survival (OS) of Japanese real-world DLBCL patients who achieved EFS24 with that of Japanese GP.

[Methods]
In this multicenter analysis, a total of 2182 patients with newly diagnosed DLBCL were retrospectively evaluated. Double hit lymphoma, primary DLBCL of the CNS, T-cell/histiocyte-rich large B-cell lymphoma, primary mediastinal large B-cell lymphoma, intravascular large B-cell lymphoma and transformed lymphomas were excluded in this analysis. All patients were treated with rituximab and anthracycline-based immunochemotherapy (ICT) as their initial therapy between 2008 and 2018. EFS was defined as time from initial ICT until relapse or progression, unplanned re-treatment of lymphoma after initial ICT, or death as a result of any cause. OS was defined as time from initial ICT for all patients or from achieving EFS12, EFS24, EFS48 and EFS60 for patients who achieved EFS12, EFS24, EFS48, EFS60, respectively until death from any cause. Each OS was compared with matched Japanese GP data.

[Results]
The median age was 71 years (range 19 to 99 years) and 53 % were male. At a median follow-up of 3.4 years, 976 patients (45%) had an event and 651 (30%) had died. The percentage of estimated survival of EFS12, EFS24, EFS48 and EFS60 was 72.9 %, 64.5 %, 55.7 % and 53.2 %, respectively. At induction of ICT, patients had a significantly decreased survival compared with matched Japanese GP, with a standardized mortality ratio (SMR) of 3.50 (95% CI 3.24 to 3.78; p < .0001; Fig 1). Although survival improved as patients remained in disease-free state, the SMR was still significant when patients achieved EFS12 (SMR 2.32; 95% CI 2.07 to 2.39; p < .0001) and EFS24 (SMR 1.71; 95% CI 1.24 to 1.86; p < .0001; Fig 1). Patients had an OS equivalent to that of the matched Japanese GP only after achieving an EFS48 (SMR 1.34; 95% CI 0.98 to 1.67; p = .0679) and EFS60 (SMR 1.27; 95% CI 0.88 to 1.68; p = .2146; Fig 1). Time-dependent ROC analysis revealed that appropriate threshold divided the patients who achieved EFS48 or not were clinical stage (CS) ≥ 3, performance status (PS) ≥ 2, and higher than 1.66 times of upper normal limit of serum level of soluble interleukin-2 receptor (sIL-2R). The model interactions in multivariate analysis excluding age and sex also demonstrated that these three factors were independent prognostic factors for achieving EFS48 (CS; p = 4.32E-8, PS; p = 1.81E-11, sIL-2R; p = 1.21E-12; Fig 2). Based on these results, we constructed prognostic index that could make a prediction of achieving EFS48, named North Japan Hematology Study Group - Prognostic Index (NJHSG-PI). This novel prognostic score successfully divided prognosis in this cohort into 4 groups; EFS and OS at 48 months from initial ICT for Low (1 point), Low-Intermediate (2 points), High-intermediate (3 points) and High risk group (4 points) were 83.4 %, 62.7 %, 46.4 % and 28.2 % (p <.0001) and 92.5 %, 79.4 %, 66.9 % and 44.5 % (p < .0001; Fig 2), respectively. Notably, patients in Low risk group achieving EFS12 had an OS equivalent to that of matched Japanese GP (SMR 1.13; 95% CI 0.81 to 1.52; p = .4714; Fig.1). For patients in other risk groups, OS was found to be equivalent to matched Japanese GPs only after achieving EFS48 or EFS60 instead of EFS24 (Fig.1).

[Conclusion]
For real-world Japanese DLBCL patients, EFS48 may be a suitable milestone instead of EFS24. NJHSG-PI using only three factors easily and effectively predict achieving EFS48. Especially, for newly DLBCL patients with early CS, good PS and low sIL-2R, 12 months after ICT could be appropriate as follow-up period.