A population level analysis of second hematological malignancies in chronic lymphocytic leukemia/small lymphocytic lymphoma survivors in the era of targeted therapies

With improvement in survival after chronic lymphocytic leukemia (CLL) diagnosis, the real‐world burden of second hematological malignancies (SHM) has not been comprehensively assessed in recent era. We analyzed risk, incidence, and outcomes of SHM in CLL patients between 2000 and 2019 using SEER database. CLL patients had greater risk for hematological malignancies than general population [SIR, standardized incidence ratio (95% CI):2.58 (2.46–2.70); p < 0.05]. The risk for subsequent lymphoma increased by 1.75 folds in 2015–2019 compared to 2000–2004. The duration, after CLL diagnosis, of maximum risk for SHM decreased as 60–119 months for time‐period 2000–2004, 6–11 months for 2005–2009 to 2–5 months for 2010–2014 and 2015–2019. Incidence of SHM was 2.5% in CLL survivors (1736/70,346) with lymphoid SHM being more common than myeloid SHM, and DLBCL being the most common pathology (n = 610, 35% of all SHM). Male sex, age ≤65 years at CLL diagnosis, and chemotherapy treatment were associated with higher risk for SHM. The median gap between CLL and SHM diagnoses was 46 months. The median survival for de‐novo‐AML, t‐MN, CML, and aggressive NHL was 63, 86, 95, and 96 months respectively. Although SHM remains rare, there is increased risk in recent era, likely due to improved survival in CLL patients, necessitating active surveillance strategies.


| INTRODUCTION
Chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL) is the most common chronic leukemia with an annual incidence of about 19,000 cases in the United States (US). 1 Based on the 2012-2014 surveillance, epidemiology and end results (SEER) data, average lifetime risk of getting CLL is about 1 in 175 (0.6%). 2 The survival of patients with CLL has significantly improved over the past 2 decades with the introduction of novel agents into clinical practice. 3,4The 5-year age-adjusted survival rate has improved from 73. 7% (1985-1989) to 89.4% (2010-2014). 36][7][8] Prior studies among non-US population have reported variable incidence of second hematologic malignancies (SHM) among CLL survivors.The most common SHM in CLL patients is Richter's transformation (RT), which is typically diffuse large B-cell lymphoma (DLBCL), although other types of lymphoma including Hodgkin lymphoma have been described. 5,9,10The Norway Cancer registry reported an incidence of 4% of SHM in CLL survivors (103/2631) between years 2003-2012, after a median follow-up of 6.6 years.SHM was more common among males, DLBCL being the most common subtype (65/103, 63%).
Notably, the median survival after SHM was significantly lower in patients who received prior treatment for CLL. 11An Australian study reported that 12% (61/517) of CLL patients (1981-2020)   developed SHM (predominantly RT to DLBCL) after a median follow-up of 11-year 12 Recently, population-based study from Netherland, reported 42% higher risk of any SHM in CLL survivors, particularly 175% greater risk of AML, compared to general population. 13In the US, studies on incidence and outcomes of second malignancies among CLL survivors are mostly limited to periods of conventional chemotherapies or chemo-immunotherapies. 7,14,15It has been noted that there is an increased incidence of second malignancies (particularly MDS and AML) with treatments using fludarabine. 7,12,14,16vel therapies have been used in routine clinical practice for almost a decade.There is risk of potentially deleterious effects on immune system with their usage which contributes to risk for SHM.
Recently, a study reported Richter syndrome and MDS/AML in 58 and 4 patients, respectively, after median follow-up of 44 months, among 691 CLL-survivors treated with BTK-inhibitors (BTKi) during 2009-2017. 17Another single center study reported higher cumulative incidence of cutaneous malignancies in patients treated with targeted therapies. 18wever, trends of different subcategories of SHM with focus on era of targeted therapies (TT) have not been comprehensively reported.The aim of this study was to assess factors influencing the risk, incidence, and outcomes of SHM, in US-population diagnosed with CLL, during 2000-2019, using the SEER database.We emphasized trends in era of oral targeted agents (including BTKi, PI3K inhibitors (PI3Ki), and BCL-2 inhibitors (BCL2i)), after approval of ibrutinib (BTKi) in year 2014.This is of particular interest for optimal utilization of healthcare resources in developing screening strategies for SHM.

| Data source
The SEER program of the National Cancer Institute (NCI) is an authoritative source of information on cancer trends in US-population and it is updated annually. 19For protecting patient privacy, only deidentified SEER data can be accessed for research purposes, after completing a research data use agreement with the NCI.We extracted data for CLL patients (ICD-O-3: 9823/3) from 17-registries-researchplus-datasets using SEER*Stat, version 8.4.0.1 multiple-primary standardized-incidence-ratio (SIR) tool. 20Data pertaining to patient's demographic profile, year of diagnosis of both the CLL and SHMs, chemotherapy treatment (as yes or no/unknown), types of SHMs, and survival status were abstracted.An approval from ethics committee was not required for this study.

| Study population
We selected adult patients (age ≥20 years) with an index diagnosis of CLL between 2000 and 2019.A subsequent diagnosis of SHM was identified with ICD-O-3/WHO 2008 site codes.We excluded SHM diagnosed within 2 months from CLL, to avoid ascertainment bias, as described in prior studies. 21We excluded patients diagnosed on autopsy or from death certificate, or with unknown survival duration.

| Standardized incidence Ratio (SIR)
The risk of hematological malignancies (HM) among CLL survivors relative to the general US population was estimated using the

| Statistical analysis
Statistical analyses were performed using GraphPad Prism version 9.5.0 for Windows, GraphPad Software, San Diego, California USA.
Descriptive statistics were calculated for patient demographics.The reverse Kaplan-Meier method was used to calculate median followup duration.Chi-squared-test was used to calculate difference between categorical variables.Unpaired-t-test was used to compare continuous variables.Kaplan-Meier method and log-rank-test were used to estimate overall survival (OS) and compare OS between subgroups, respectively.A multivariate logistic regression was performed to assess variables (age, gender, race, time-period of diagnosis and chemotherapy status) predicting development of SHM in CLL patients.A two-sided p-value <0.05 was considered statistically significant.

| Patient and SHM characteristics
We found 70,346 patients diagnosed with CLL during 2000-2019.

| Risk and time trend of SHM among CLL survivors
The overall incidence rate of SHM was 4.49 per-1000 person-years (py) follow up of CLL survivors which sequentially increased for 5year periods between 2000 and 2019 (Figure 1).There was 158% higher risk of HM in CLL survivors compared to age, gender, race and period-matched general population [SIR (95%CI): 2.58 (2.46-2.70),p < 0.05] during 2000-2019.Overall, maximum risk was observed during initial 2-5-month after CLL diagnosis, and it declined for consecutive increase in latency duration (Table 2).When further stratified by 5-year time-periods, the latency period, after CLL diagnosis, having maximum risk of SHM declined from 60 to 119 months in 2000-2004 to 2-5 months in 2015-2019 (Table 2).
We observed an increasing trend of overall SIR for all SHM through 5-year-periods.With regards to type of SHM, the risk of developing subsequent lymphoma was significantly higher through each period and increased by 1.75 folds in 2015-2019 compared to 2000-2004 (Table 2).The risk of leukemia was significantly elevated during years 2005-2009 and 2010-2014 (Table 2).On the other hand, the risk of myeloma did not significantly differ in each period.Overall, maximum risk of subsequent lymphoma was during 6-11 months after CLL diagnosis, while for subsequent myeloma, during initial 2-5 months (Table 2).
We assessed trend of different histologic sub-types of lymphoid SHM for consecutive 5-year-periods of CLL diagnosis.Higher proportion of ANHL was observed among patients with more recent CLL diagnosis-period (Table 1).There was no significant difference in the incidence of histologic subtypes of myeloid SHM, though majority of t-MN (32/63, 50.8%) were observed in patients with CLL diagnosis during 2010-2014.

| Patient-related factors
The overall proportions of lymphoid and myeloid SHM were similar between patients aged ≤65 years (younger) and aged >65 years (older) at CLL diagnosis.Upon stratifying for histologic subtypes, there was higher proportion of t-MN in those with younger age while of INHL, PCM, CML and AML in those with older age at CLL diagnosis (Table 1).Stratification by sex (female vs. male), or race (white, black vs. others), did not reveal significant difference in proportions of each lineage of SHM (lymphoid vs. myeloid) or between the major histologic sub-types of each lineage (Table 1).

| Factors predicting development of SHM
We compared data of CLL patients without history of another malignancy (n = 52,321) with CLL patients who developed SHM using multivariate regression analysis.We assessed the odds ratio (OR) for developing SHM in CLL patients based on age at CLL diagnosis, sex, race, chemotherapy treatment status and time period of diagnosis.
The odds of developing SHM [OR (95% CI)] was 20% lower with age >65 years versus ≤65 years at CLL diagnosis [0.8 (0.73-0.88), p < 0.001], which could be due to a shorter survival or follow-up in the older age group.Additionally, it was higher in males, and those who received chemotherapy (Figure 2).

| DISCUSSION
In this large population-based study, with a long-term follow-up, we found 2.5% cumulative incidence of SHM and an incidence-rate (IR) of 4.49 per-1000 py follow-up.Prior studies conducted in smaller cohort size have reported similar findings of incidence ranging 1.3%-5.2%and IR per-1000 py ranging 1.1-6.4. 7,11,12,17,23,24 found 158% higher risk of hematological malignancies among CLL survivors compared to general US population during 2000-2019.
The risk is higher compared to that reported in prior studies, perhaps due to relatively recent period of CLL diagnosis (year 2000 onwards) included in our study than the prior studies (year 1973 onwards). 13,21ere has been a paradigm shift in treatment for CLL in this period and there is a growing concern for associated increment in second aggressive malignancies.During this period studies have reported higher risk of SHM in CLL survivors, partly attributable to fludarabine-based chemotherapies. 12,14,16,17With advancements in    2].
precision medicine, BTKi (ibrutinib) was approved as one of the initial oral targeted therapies (TT) for CLL in February 2014, 25 followed by approval of other agents (PI3Ki and BCL2i).Since then, single center studies have also reported risk of second malignancies in patients treated with TT. 17,18 This elevated risk for development of a subtype of second cancers can be explained by ongoing immune dysfunction in CLL survivors, potentially magnified by therapy-induced DNA damage. 16,26Additionally, immune suppression by novel therapies, and impaired immunosurveillance of clonal hematopoiesis, likely aggravated after treatment with chemotherapy, have also been hypothesized for development of second hematological cancers. 27,28 assessed the association between 5-year-periods of CLL A multivariate analysis revealed increased risk of SHM in CLL patients with male sex, younger age at diagnosis, and exposure to chemotherapy, consistent with prior studies. 13,16,21In alignment with prior studies, we noted development of SHM significantly reduces OS in CLL patients (OS: 99 vs. 127 months). 11,12,29Among myeloid SHM, CML (95 months) had the best while de-novo AML (63 months) had the poorest median OS after CLL diagnosis, a finding consistent with prior studies. 12,29e main strengths of our study are a) large sample size of CLL MISHRA ET AL.SIR, defined as the ratio of observed to expected number of events.The referent rate file on expected HM in the general population (based on US census for year 2000) is incorporated in SEER stat, version 8.4.0.1.It has been calculated by multiplying the incidence rates specific for gender, race, attained age, and calendaryear by the specific person-years at risk.Poisson distribution of observed SHM was assumed for calculation of the 95% confidence intervals (CIs) and "p" value for SIR.For comparison of SIR across different time periods, ICD-O-3/WHO 2008 site codes for lymphoma, myeloma and leukemia were used (Supplementary Table

F I G U R E 1
Incidence rate of second hematological malignancies (SHM) per 1000person-years among CLL survivors.888 -MISHRA ET AL.

F I G U R E 2
Multivariate analysis of factors predicting development of second hematological malignancies after CLL diagnosis.F I G U R E 3 (A) Survival after chronic lymphocytic leukemia (CLL) diagnosis, compared between patients with second hematological malignancies (SHM) versus patients with CLL only.(B) Survival after chronic lymphocytic leukemia (CLL) diagnosis, compared between patients with lymphoid versus myeloid second hematological malignancies (SHM).(C) Survival after chronic lymphocytic leukemia (CLL) diagnosis, stratified by subcategories of lymphoid second hematological malignancies.(D) Survival after CLL diagnosis, stratified by subcategories of myeloid second hematological malignancies [AML = acute myeloid leukemia, CML = chronic myeloid leukemia and t-MN = therapy related myeloid neoplasm, as categorized in Supplementary Table survivors identified via population-based registry data, which eliminated the selection bias seen in single institution-based studies, b) quantitative assessment of incidence, risks, and survival by SHM subtype.We limited our analysis to patients diagnosed with CLL after 2000 to assess the trends in recent time-periods with paradigm shift in CLL treatment landscape.Additional limitations are a)underascertainment of second malignancy risks because of patient migration outside SEER program, b)using time-period as surrogate for treatment with targeted therapies, may have attenuated our results, c)shorter follow up of patients who were diagnosed with CLL in recent time-period.6| CONCLUSIONIn conclusion, SHM are rare with an incidence of 2.5% among CLL survivors, but these significantly lower overall survival for CLL patients.There is progressive risk of subsequent lymphoma, with increasing proportion of aggressive NHL, in CLL patients diagnosed during recent time-period.The decline of latency period with maximum risk of SHM to a shorter period, in recent era of targeted therapies, perhaps reflects increased medical awareness and improved diagnostics in clinical oncology in recent era.It further necessitates effective surveillance for SHM, starting early after CLL diagnosis.Future prospective studies assessing the impact of targeted therapy on SHM are required.AUTHOR CONTRIBUTIONRahul Mishra was responsible for concept and design, data extraction and cleaning, statistical analysis, interpreting results, and writing the manuscript.Danai Dima and Agrima Mian were responsible for study design and writing the manuscript.Sumukh A Kumar, Alankrita Taneja and Rahul Karna were responsible for making results table, and proofreading the manuscript.Paolo F Caimi, and Deepa Jagadeesh were responsible for conceptualization, supervision, providing feedback and manuscript editing.Robert Dean and Brian T Hill provided feedback on the manuscript.All authors reviewed and approved the final version of manuscript.MISHRA ET AL.