In this comprehensive US population-level comparative analysis of patients with transformed iNHL and de novo DLBCL, we report several important observations. First, the incidence of HT was highest in SMZL, followed by FL, and lowest in EMZL. Second, 5-year RS, OS, and LSS was better in patients with de novo DLBCL compared to transformed iNHL subtypes. Among the patients with iNHL, t-LPL/WM had the worse survival rates relative to the other subtypes. Third, early HT and receipt of prior chemotherapy was associated with inferior survival for patients with t-FL but not other iNHL subtypes. Our findings add to the growing body of literature on the epidemiology of HT of iNHLs and will serve as a guide to explore further research in tailoring the management based on the transformed iNHL subtype and the timing of HT.

Notably, for t-FL, our results are lower than those reported in other studies conducted after the introduction of rituximab, with studies reporting a cumulative incidence of HT in FL of 10.7% at 5 years, between 4.1 to 15.3% at 6 years, and 7.7 to 8% at 10 years [2,3,4,5, 7]. However, in our analysis, we report a lower rate of HT (5.5%) at a median follow-up of 7.75 years, a finding closer to that recently published study by Florindez et al. [6] These discrepancies may be attributed to variations in population characteristics, geographic factors influencing treatment decisions, study design (including rituximab treatment rates), and assessment methodologies. The cumulative incidence of HT in MZL at 10 years ranged from 2.95 to 8.4% across various studies, with SMZL having the highest incidence and EMZL with the lowest rates of HT [6, 13, 14, 17, 19], which is in line with our study. Moreover, the rate of HT in LPL/WM in our study was similar to the previously reported 2.4% at both 5 and 10 years [16, 18]. Our analysis, which reports a 2.2% cumulative incidence over a median of 7.67 years, bridges these two time points. The discrepancies among these findings could be attributed to several factors, notably the rarity of WM/LPL, the underlying treatment strategies, as well as the power of a population-level analysis compared to smaller cohorts.

The differences in the incidence of HT across different lymphoma subtypes may be attributed to distinct molecular mechanisms that drive the transformation in each. For example, in FL, genome-wide studies have provided insight into the molecular alterations that contribute to HT. T-FL has been shown to harbor a high mutational burden, particularly due to increased copy number aberrations, mutational structural rearrangements, and somatic hypermutation of target genes, thus increasing genomic complexity and instability [20, 21]. There is emerging data regarding the clinicopathological and molecular characteristics of MZL and WM leading to HT. Studies have shown that the presence of monoclonal protein [22], high Ki-67% [23], and lack of achievement of complete response to first-line treatment were associated with an increased risk of transformation of MZL [17]. Additionally, comprehensive molecular analyses identified NF-kB signaling genes (TNFAIP3 and KMT2D), NOTCH/2 pathway genes, KLF2, and TP53 as the most commonly altered genes in t-SMZL, with KLF2 and complex chromosomal structures being associated with inferior survival [24,25,26]. In WM, the presence of wild-type MYD88 has been shown to confer a higher risk of HT [18, 27, 28] associated with carrying a higher risk of mutation in several genes that contribute to NF-kB signaling, similar to the genetic profile of de novo DLBCL [29, 30].

In contrast to two non-population level studies reporting no difference in overall survival in transformed versus de novo DLBCL [15, 31], our analysis showed inferior survival in transformed iNHL compared to de novo DLBCL. Our adjusted analysis supports the findings of recent studies showing a significantly higher risk of mortality in patients with t-FL [6, 12] compared to de novo DLBCL. However, in contrast to the recent study by Florindez et al. [6], this association was not statistically significant in any of the t-MZL subtypes. To our knowledge, we are the first to report an increased risk of mortality with t-LPL/WM compared to de novo DLBCL. Although the exact reasons are unclear for this trend, the poorer prognosis associated with t-FL and t-LPL/WM may be driven by the presence of high-risk molecular features that drive transformation, in addition to the influence of prior therapy.

We found that patients with FL had the highest frequency of early HT, with 47% of patients experiencing HT within 2 years of initial diagnosis, followed by t-LPL/WM (42%) and t-MZL (36%). We noted that the occurrence of early HT was associated with inferior OS in patients with t-FL, with a 34% increase in risk of mortality compared to late HT, which is in line with the prior studies [5,6,7]. However, we did not see any association between the timing of HT on outcomes in t-MZL and t-LPL/WM. While the impact of timing of HT on outcomes in t-MZL has been shown in prior work [6], this is the first study to evaluate the impact of timing of HT on outcomes in patients with t-LPL/WM. This is important information that can be useful while counseling patients.

Receipt of chemotherapy prior to HT was highest for t-FL, followed by t-LPL/WM and t-MZL. In line with published data prior treatment was associated with inferior survival for t-FL [6, 8, 9], however this finding was not significant with other subtypes. One possible explanation for this observation in FL is that prior exposure to chemotherapy may select resistant clones that contribute to the development of transformation and treatment refractoriness [1, 32].

Our study is limited by a lack of certain clinicopathologic information at the patient level, such as International Prognostic Index scores, cytogenetic abnormalities, and molecular alterations. These data are not consistently reported to cancer registries in the United States and are not included in the SEER database. While there was information pertaining to the receipt of prior systemic therapy, the exact details of the treatment regimen, including prior anthracycline exposure, are not available in the SEER cancer registry. Hence, we could not match the two groups (de novo DLCBL and transformed lymphoma) for the type of systemic therapy and number of lines of therapy.

In conclusion, our findings provide a comprehensive insight into the incidence of transformation and prognosis of transformed iNHLs. To our knowledge, this is the first comparative study to show that the outcomes of patients with t-LPL/WM were inferior compared to de novo DLBCL. We also identified different rates of transformation for MZL subtypes, with the highest incidence in SMZL. This emphasizes the need to re-biopsy when SMZL patients present with the progression of the disease to rule out HT. We found that patients with transformed iNHLs have poor survival compared to de novo DLBCL, which is important for counseling patients in the clinic. Additionally, the poor prognosis associated with early HT and receipt of prior chemotherapy in patients with t-FL underscores the need for early institution of experimental therapies in this high-risk subgroup.