Abstract
BACKGROUND:
Previous studies have shown an increased risk of secondary primary malignancies (SPMs) after diffuse large B-cell lymphoma (DLBCL) treatment. Whether stage of DLBCL at diagnosis affects the subtypes of SPMs that occur has not been previously described.
METHODS:
The Surveillance, Epidemiology, and End Results database was queried for patients aged >18 years diagnosed with primary DLBCL from 1973 to 2010 and categorized by early stage (ES) (stage I-II) or advanced stage (AS) (stage III-IV) disease. Differences in overall and location-specific SPM incidence by stage and time since diagnosis were assessed in 5-year intervals using a Fine-Gray hazards model. Overall survival was compared using the log-rank test. A Cox proportional hazards model was used to assess differences in survival.
RESULTS:
In total, 26,038 patients with DLBCL were identified, including 14,724 with ES and 11,314 with AS disease. The median follow-up was 13.3 years. Overall, 13.0% of patients developed SPM, with a higher but nonsignificantly increased risk of SPM development in those who had ES disease compared with those who had AS disease (14% vs 11.6%; P = .14). During the first 5 years after diagnosis, patients who had ES disease had a higher risk of SPM than those who had AS disease, specifically colorectal, pancreas, breast, and prostate SPMs. During the period from 10 to 15 years after diagnosis, patients who had AS disease had a higher risk of SPM than those who had ES disease, specifically hematologic SPMs. Development of SPM was found to significantly increase the risk of death regardless of stage at diagnosis.
CONCLUSIONS:
In this large, population-based study, distinctly different subtypes and temporal patterns of SPM development were identified based on stage of DLBCL at diagnosis. The current study merits consideration of tailored site-specific and time-specific surveillance for patients with DLBCL according to stage and time interval since diagnosis.
Keywords: diffuse large B-cell lymphoma, non-Hodgkin lymphoma, second primary neoplasms, survivorship
INTRODUCTION
Among all types of cancer, survivors of non-Hodgkin lymphoma (NHL) are at higher risk of developing a second cancer in their lifetimes.1,2 These second cancers, known as secondary primary malignancies (SPMs), are an emerging dilemma in the current landscape of lymphoma treatment, as surveillance strategies are not well defined. The most common type of NHL is diffuse large B-cell lymphoma (DLBCL), for which survival has improved considerably since the advent of rituximab, with a cure for most patients who have early stage (ES) disease and for 50% of patients who have advanced-stage (AS) disease.3 With increasing survival after DLBCL, the elucidation of which patient and disease characteristics predict an increased risk of SPM development would meaningfully contribute to survivorship management.
The etiology of SPM after DLBCL is multifactorial and has been suggested to be caused by a combination of exposure to chemotherapy, radiation, rituximab and associated immunosuppression, chronic infections, lifestyle practices, demographics, and genetic susceptibility.2,4–6 The genetic landscape of DLBCL is complex, with a heterogeneous presentation that differs between ES and AS disease.7,8 Previous literature has demonstrated an increased risk of late relapses in patients who have ES compared with those who have AS DLBCL, which suggests a unique biology of ES DLBCL.8–10
Given this potential difference in biology, it is possible that the subtypes of SPM that develop after ES versus AS DLBCL differ as well, which has important clinical implications on the long-term follow-up of DLBCL survivors. The effect of DLBCL stage on SPM risk has not been specifically studied, and current guidelines on survivorship do not provide specific direction on SPM screening in patients treated for DLBCL. We designed the current study to evaluate whether DLBCL stage at diagnosis is associated with specific subtypes of SPM and whether stage at diagnosis affects the time until development of various subtypes of SPM.
MATERIALS AND METHODS
The Surveillance, Epidemiology, and End Results (SEER) 9 Regs Research Data, Nov 2016 Sub (1973–2014) for Standardized Mortality Ratios (SMRs) database were used to identify patients aged ≥18 years who were diagnosed with a first primary DLBCL (lymphoma subtype recode/World Health Organization, 2008: 2[a]2.3.1 DLBCL, not otherwise specified) from 1973 to 2010. Stage at diagnosis of first primary DLBCL was categorized into ES for stage I and II disease or AS for stage III and IV disease. Those with unknown stage information were excluded from the study. SPM was defined in this study as a second malignancy occurring at least 6 months after a first primary DLBCL.11 To account for the treatment effect of rituximab, an indicator variable was used to identify patients who were diagnosed in the prerituximab versus postrituximab era. The year 2001 was used to mark the start of the rituximab era.5 All analyses were performed using R version 3.4.112 and SAS version 9.4.13 For individuals with multiple SPMs, only information from their first diagnosed SPM was used. SPMs in the SEER database are categorized using the third edition of the International Classification of Diseases for Oncology (ICD-O-3) code, which was re-categorized into 20 different site locations, as described in Supporting Table 1.
A Fine-Gray model was used to evaluate the relationships between age, sex, race, stage, extranodal disease, surgery, radiation therapy, rituximab era, and the incidence of SPM over the entire study period from 1973 to 2010, while accounting for death as a competing risk and adjusting for sex, rituximab, age at diagnosis, and race.14,15 The Fine-Gray model was conducted in R using the Cmprsk package.16 P values and 95% CIs were adjusted to account for multiple comparisons using Bonferroni correction. We further explored SPM incidence in 5-year intervals using similar techniques to determine whether these associations varied over time. This exploratory analysis was not adjusted for multiple comparisons.
A Cox proportional hazards model was used to examine whether survival differed by SPM development, stage of DLBCL at diagnosis, and diagnosis in the prerituximab or postrituximab era, while adjusting for age, sex, and race. To avoid survival bias and account for the variability in the time of SPM development, SPM was modeled as time-dependent covariate in the Cox proportional hazards model using PHREG and the counting process method in SAS. A 3-way interaction among these variables was included in the model to assess the dependence of these variables on one another.
RESULTS
Patient and Disease Characteristics
Our SEER query identified 26,038 adults who were diagnosed with DLBCL between 1973 and 2010, of whom 14,724 had ES disease and 11,314 had AS disease. Patient and disease characteristics for the ES and AS cohorts are presented in Table 1. The median age at diagnosis was 63 years, and there was a slightly higher incidence in males. Among the patients identified, 14,312 (ES cohort, n = 8399; AS cohort, n = 5913) were diagnosed in the prerituximab era, and 11,726 (ES cohort, n = 6325, AS cohort, n = 5401) were diagnosed in the postrituximab era.
TABLE 1.
Overview of Individuals From the Surveillance, Epidemiology, and End Results Database With a Diffuse Large B-Cell Lymphoma As Their Primary Malignancy
| Variable | Level | Stage Type: No. of Patients (%)a | P | |
|---|---|---|---|---|
| ES, n = 14724 | AS, n = 11314 | |||
| Age at diagnosis: Median [IQR], y | 63.00 [49.00–74.00] | 63.00 [50.00–73.00] | .017 | |
| Age by category, y | 18–59 | 6295 (42.8) | 4895 (43.3) | .416 |
| ≥60 | 8429 (57.2) | 6419 (56.7) | ||
| Sex | Female | 6898 (46.8) | 5041 (44.6) | <.001 |
| Male | 7826 (53.2) | 6273 (55.4) | ||
| Race | American Indian/Alaska Native | 69 (0.5) | 67 (0.6) | <.001 |
| Asian or Pacific Islander | 1217 (8.3) | 681 (6.0) | ||
| Black | 834 (5.7) | 890 (7.9) | ||
| White | 12,604 (85.6) | 9676 (85.5) | ||
| Year of diagnosis | <1992 | 3304 (22.4) | 2320 (20.5) | <.001 |
| >2001 | 5677 (38.6) | 4905 (43.4) | ||
| 1992–2001 | 5743 (39.0) | 4089 (36.1) | ||
| SEER registry | Metropolitan Atlanta, 1975+ | 1112 (7.6) | 852 (7.5) | <.001 |
| Connecticut, 1973+ | 2336 (15.9) | 1553 (13.7) | ||
| Metropolitan Detroit, 1973+ | 1889 (12.8) | 1806 (16.0) | ||
| Hawaii, 1973+ | 626 (4.3) | 396 (3.5) | ||
| Iowa, 1973+ | 2102 (14.3) | 1626 (14.4) | ||
| New Mexico, 1973+ | 821 (5.6) | 479 (4.2) | ||
| San Francisco-Oakland SMSA, 1973+ | 2632 (17.9) | 2022 (17.9) | ||
| Seattle-Puget Sound, 1974+ | 2161 (14.7) | 1844 (16.3) | ||
| Utah, 1973+ | 1045 (7.1) | 736 (6.5) | ||
| Cause of death | Alive | 6094 (41.4) | 3808 (33.7) | <.001 |
| Not cancer-related | 3138 (21.3) | 1824 (16.1) | ||
| Unknown | 509 (3.5) | 500 (4.4) | ||
| Cancer-related | 4983 (33.8) | 5182 (45.8) | ||
| DLBCL primary site location) | Central nervous system | 707 (4.8) | 149 (1.3) | <.001 |
| Head and neck | 1497 (10.2) | 281 (2.5) | ||
| Lymph nodes | 7184 (48.8) | 9035 (79.9) | ||
| Foregut and midgut | 1723 (11.7) | 553 (4.9) | ||
| Colorectal | 427 (2.9) | 124 (1.1) | ||
| Hepatobiliary | 101 (0.7) | 73 (0.6) | ||
| Pancreas | 69 (0.5) | 38 (0.3) | ||
| Pulmonary | 215 (1.5) | 130 (1.1) | ||
| Thyroid and thymus | 403 (2.7) | 57 (0.5) | ||
| Heart | 17 (0.1) | 6 (0.1) | ||
| Mediastinum | 120 (0.8) | 31 (0.3) | ||
| Skeletal | 377 (2.6) | 167 (1.5) | ||
| Skin | 427 (2.9) | 71 (0.6) | ||
| Breast | 216 (1.5) | 45 (0.4) | ||
| Female genital | 102 (0.7) | 52 (0.5) | ||
| Urinary | 122 (0.8) | 65 (0.6) | ||
| Male genital | 346 (2.3) | 77 (0.7) | ||
| Soft tissue | 495 (3.4) | 163 (1.4) | ||
| Hematologic | 167 (1.1) | 166 (1.5) | ||
| Unknown | 9 (0.1) | 31 (0.3) | ||
| Extranodal disease | No | 7184 (48.8) | 9035 (79.9) | <.001 |
| Yes | 7540 (51.2) | 2279 (20.1) | ||
| Radiation therapy | No/unknown | 8848 (60.1) | 9256 (81.8) | <.001 |
| Yes | 5876 (39.9) | 2058 (18.2) | ||
| Surgery | No | 5527 (37.5) | 5335 (47.2) | <.001 |
| Yes | 9197 (62.5) | 5979 (52.8) | ||
| SPM | No | 12,661 (86.0) | 9999 (88.4) | <.001 |
| Yes | 2063 (14.0) | 1315 (11.6) | ||
| SPM site location | None | 12,661 (86.0) | 9999 (88.4) | <.001 |
| Central nervous system | 39 (0.3) | 27 (0.2) | ||
| Head and neck | 85 (0.6) | 65 (0.6) | ||
| Lymph nodes | 85 (0.6) | 81 (0.7) | ||
| Foregut and midgut | 80 (0.5) | 35 (0.3) | ||
| Colorectal | 232 (1.6) | 117 (1.0) | ||
| Hepatobiliary | 39 (0.3) | 31 (0.3) | ||
| Pancreas | 39 (0.3) | 30 (0.3) | ||
| Pulmonary | 253 (1.7) | 164 (1.4) | ||
| Thyroid and thymus | 39 (0.3) | 23 (0.2) | ||
| Mediastinum | 1 (0.0) | 0 (0.0) | ||
| Skeletal | 7 (0.0) | 5 (0.0) | ||
| Skin | 163 (1.1) | 115 (1.0) | ||
| Breast | 236 (1.6) | 108 (1.0) | ||
| Female genital | 80 (0.5) | 63 (0.6) | ||
| Urinary | 184 (1.2) | 115 (1.0) | ||
| Male genital | 297 (2.0) | 157 (1.4) | ||
| Soft tissue | 17 (0.1) | 11 (0.1) | ||
| Hematologic | 151 (1.0) | 136 (1.2) | ||
| Unknown | 36 (0.2) | 32 (0.3) | ||
| Survival: Median [IQR], mo | 85.00 [29.00–151.00] | 57.00 [14.00–118.00] | <.001 | |
| Rituximab era: Year of Dx >2001 | No | 8399 (57.0) | 5913 (52.3) | <.001 |
| Yes | 6325 (43.0) | 5401 (47.7) | ||
Abbreviations: AS, advanced stage; DLBCL, diffuse large B-cell lymphoma; Dx, diagnosis; ES, early stage; IQR, interquartile range; SEER, Surveillance, Epidemiology, and End Results program of the National Cancer Institute; SMSA, standard metropolitan statistical area; SPM, secondary primary malignancy.
The cohort has been stratified by DLBCL stage at the time of diagnosis into early (stage ≤2) and advanced (stage >2) disease.
Incidence of SPM
Overall, 13.0% of the study population developed an SPM. The most common SPM subtypes were hematologic (leukemia and lymphoma), male genital, pulmonary, colorectal, breast, urinary, and skin. The median time to SPM diagnosis was 62 months (interquartile range [IQR], 25–120 months) for the entire cohort, and it was 65 months (IQR, 26–126 months) and 59 months (IQR, 22–112 months) for the ES and AS cohorts, respectively. By using the reverse Kaplan-Meier method, the median follow-up was 13.3 years (95% CI, 13.1–13.4 years), with a median follow-up that was slightly longer for the ES cohort than for the AS cohort at 13.9 years (95% CI, 13.7–14.1 years) and 12.3 years (95% CI, 12–12.5 years), respectively.
Patients in the ES cohort had a higher incidence of SPM than those in the AS cohort (14.0% vs 11.6%, respectively) (Table 1). The 30-year cumulative incidence of SPM was 21.26% in the ES cohort and 17.46% in the AS cohort (P = .135), as depicted in Figure 1. In our study, 51.2% of the ES cohort and 20.1% of the AS cohort had extranodal disease (Table 1). Patients with extranodal disease at diagnosis had a significantly higher risk of SPM development (hazard ratio [HR], 1.15; 95% CI, 1.01–1.32; P = .018) that was independent of stage, as depicted in Figure 2.
Figure 1.
The cumulative incidence of secondary primary malignancy (SPM) stratified by stage of diffuse large B-cell lymphoma at diagnosis (advanced vs early) is illustrated. P values and CIs for the hazard ratio (HR) have be adjusted using Bonferroni correction for multiple comparisons. The risk table below the graph reports 1) the number at risk of SPM and death at the beginning of each time interval; and 2) in parentheses, the number of SPM events within a time interval. Note that the scale on the y-axis spans from 0% to 20%.
Figure 2.
Cumulative incidence of secondary primary malignancy (SPM) stratified by the presence of extranodal disease (END) status (yes vs no) is illustrated. P values and CIs for the hazard ratio (HR) have be adjusted using Bonferroni correction for multiple comparisons. The risk table below the graph reports 1) the number at risk of SPM and death at the beginning of each time interval; and 2) in parentheses, the number of SPM events within a time interval. Note that the scale on the y-axis spans from 0% to 20%.
RISK FACTORS FOR SPM DEVELOPMENT IN AGGREGATE ANALYSIS, 1973 TO 2010
Risk factors for SPM development over the entire study period (1973–2010), for all SPM locations in aggregate as well as for individual SPM subtypes, are presented in Tables 2 and 3. The risk of developing an SPM was 12% lower for the AS cohort compared with the ES cohort, but it did not reach statistical significance after accounting for death as a competing risk and multiple comparisons (P = .135). There was a significantly higher risk of SPM development for white race, male gender, extranodal disease, and diagnosis in the postrituximab era for all SPM locations in aggregate. Advancing age was also significantly associated with an increased risk of SPM (HR, 1.01). Radiation therapy was not associated with a significant increase in SPM risk. Diagnosis in the postrituximab era was associated with a significantly increased risk of lymph node (HR, 2.44), thymus and thyroid (HR, 2.99), and hematologic (HR, 1.99) SPMs.
TABLE 2.
Variables Used in the Fine-Gray Model to Assess Differences in the Incidence of Secondary Primary Malignancya
| Variable | Reference | Definition |
|---|---|---|
| Stage | Advanced vs early | Stage at diagnosis of DLBCL: early, stage ≤II; advanced, stage >II |
| Surgery | Yes vs no | Indicator of surgery, yes/no |
| Extranodal | Yes vs no | Indicator of extranodal disease, yes/no |
| Radiation | Yes vs no | Indicator of beam radiation, yes/no |
| Race | White vs nonwhite | Indicator of race, white/nonwhite |
| Age | Not applicable | Continuous variable of age at diagnosis |
| Sex | Male vs female | Indicator of sex, male/female |
| Rituximab | Yes vs no | Yes, diagnosed after 2001 in rituximab era |
This was assessed overall and for specific SPM sites, as detailed in Table 3.
TABLE 3.
Fine-Gray Model Results Assessing Whether Secondary Primary Malignancy Incidence Was Different Between Those Diagnosed With an Early or Advanced-Stage Diffuse Large B-Cell Lymphoma, Diagnosed in the Post-Rituximab Era, Had Extranodal Disease, Received Beam Radiation, Underwent Site-Specific Surgery, and Demographics
| SPM Siteb | Variable | HR | Corrected 95% CIa | Corrected Pa | |
|---|---|---|---|---|---|
| Lower | Upper | ||||
| Overall, N = 26,038 | Stage | 0.88 | 0.77 | 1.01 | .135 |
| Rituximab | 1.16 | 1.01 | 1.32 | .01296 | |
| Race | 1.37 | 1.13 | 1.68 | >.00001 | |
| Surgery | 1.06 | 0.93 | 1.22 | 1.00 | |
| Extranodal | 1.15 | 1.01 | 1.32 | .01755 | |
| Radiation | 1.03 | 0.90 | 1.17 | 1.00 | |
| Sex | 1.31 | 1.15 | 1.49 | >.00001 | |
| Age | 1.01 | 1.005 | 1.012 | >.00001 | |
| Central nervous system, N = 26,025 | Stage | 0.94 | 0.36 | 2.47 | 1.00 |
| Rituximab | 1.69 | 0.65 | 4.43 | 1.00 | |
| Race | 0.98 | 0.29 | 3.31 | 1.00 | |
| Surgery | 1.03 | 0.37 | 2.86 | 1.00 | |
| Extranodal | 1.20 | 0.47 | 3.09 | 1.00 | |
| Radiation | 0.95 | 0.37 | 2.46 | 1.00 | |
| Sex | 0.79 | 0.33 | 1.89 | 1.00 | |
| Age | 1.00 | 0.98 | 1.02 | 1.00 | |
| Head and neck, N = 26,025 | Stage | 1.22 | 0.67 | 2.23 | 1.00 |
| Rituximab | 0.82 | 0.42 | 1.58 | 1.00 | |
| Race | 1.44 | 0.55 | 3.81 | 1.00 | |
| Surgery | 1.28 | 0.65 | 2.53 | 1.00 | |
| Extranodal | 1.50 | 0.82 | 2.72 | 1.00 | |
| Radiation | 1.10 | 0.60 | 2.03 | 1.00 | |
| Sex | 2.06 | 1.08 | 3.94 | .008505 | |
| Age | 1.00 | 0.99 | 1.02 | 1.00 | |
| Lymph nodes, N = 26,025 | Stage | 1.15 | 0.62 | 2.12 | 1.00 |
| Rituximab | 2.44 | 1.33 | 4.46 | >.0001 | |
| Race | 1.06 | 0.48 | 2.34 | 1.00 | |
| Surgery | 1.11 | 0.61 | 2.01 | 1.00 | |
| Extranodal | 1.01 | 0.55 | 1.88 | 1.00 | |
| Radiation | 0.78 | 0.39 | 1.55 | 1.00 | |
| Sex | 1.23 | 0.69 | 2.19 | 1.00 | |
| Age | 0.99 | 0.97 | 1.00 | .3375 | |
| Foregut and midgut, N = 26,025 | Stage | 0.60 | 0.28 | 1.31 | 1.00 |
| Rituximab | 0.68 | 0.32 | 1.44 | 1.00 | |
| Race | 0.69 | 0.29 | 1.67 | 1.00 | |
| Surgery | 0.83 | 0.39 | 1.76 | 1.00 | |
| Extranodal | 1.13 | 0.55 | 2.33 | 1.00 | |
| Radiation | 0.94 | 0.45 | 1.97 | 1.00 | |
| Sex | 2.06 | 1.02 | 4.16 | .03375 | |
| Age | 1.02 | 1.00 | 1.04 | .06615 | |
| Colorectal, N = 26,025 | Stage | 0.76 | 0.50 | 1.17 | 1.00 |
| Rituximab | 0.89 | 0.59 | 1.35 | 1.00 | |
| Race | 1.08 | 0.61 | 1.91 | 1.00 | |
| Surgery | 1.04 | 0.68 | 1.57 | 1.00 | |
| Extranodal | 1.37 | 0.92 | 2.04 | .648 | |
| Radiation | 1.07 | 0.71 | 1.61 | 1.00 | |
| Sex | 1.08 | 0.74 | 1.59 | 1.00 | |
| Age | 1.01 | 1.00 | 1.02 | .04995 | |
| Hepatobiliary, N = 26,025 | Stage | 0.95 | 0.41 | 2.18 | 1.00 |
| Rituximab | 1.43 | 0.57 | 3.58 | 1.00 | |
| Race | 0.52 | 0.19 | 1.44 | 1.00 | |
| Surgery | 0.58 | 0.23 | 1.45 | 1.00 | |
| Extranodal | 1.12 | 0.45 | 2.80 | 1.00 | |
| Radiation | 0.79 | 0.30 | 2.06 | 1.00 | |
| Sex | 2.04 | 0.82 | 5.04 | .6885 | |
| Age | 1.01 | 0.98 | 1.03 | 1.00 | |
| Pancreas, N = 26,025 | Stage | 0.95 | 0.36 | 2.48 | 1.00 |
| Rituximab | 0.72 | 0.27 | 1.93 | 1.00 | |
| Surgery | 0.72 | 0.27 | 1.88 | 1.00 | |
| Extranodal | 0.73 | 0.26 | 2.09 | 1.00 | |
| Radiation | 1.11 | 0.42 | 2.89 | 1.00 | |
| Sex | 1.40 | 0.57 | 3.42 | 1.00 | |
| Age | 1.02 | 1.00 | 1.05 | .04185 | |
| Pulmonary, N = 26,025 | Stage | 0.88 | 0.59 | 1.29 | 1.00 |
| Rituximab | 0.89 | 0.61 | 1.30 | 1.00 | |
| Race | 1.09 | 0.63 | 1.88 | 1.00 | |
| Surgery | 0.95 | 0.65 | 1.39 | 1.00 | |
| Extranodal | 1.03 | 0.70 | 1.53 | 1.00 | |
| Radiation | 0.98 | 0.66 | 1.44 | 1.00 | |
| Sex | 1.11 | 0.77 | 1.59 | 1.00 | |
| Age | 1.02 | 1.01 | 1.03 | >.00001 | |
| Thyroid and thymus, N = 26,025 | Stage | 1.00 | 0.35 | 2.87 | 1.00 |
| Rituximab | 2.99 | 1.15 | 7.73 | .005535 | |
| Surgery | 1.28 | 0.48 | 3.43 | 1.00 | |
| Extranodal | 1.76 | 0.67 | 4.63 | 1.00 | |
| Radiation | 1.38 | 0.52 | 3.68 | 1.00 | |
| Sex | 0.31 | 0.11 | 0.86 | .00567 | |
| Age | 0.98 | 0.95 | 1.00 | .01755 | |
| Skin, N = 26,025 | Stage | 0.97 | 0.61 | 1.56 | 1.00 |
| Rituximab | 1.32 | 0.84 | 2.08 | 1.00 | |
| Race | 3.73 | 1.31 | 10.63 | .0010125 | |
| Surgery | 1.08 | 0.67 | 1.73 | 1.00 | |
| Extranodal | 1.08 | 0.68 | 1.72 | 1.00 | |
| Radiation | 1.20 | 0.76 | 1.90 | 1.00 | |
| Sex | 3.16 | 1.86 | 5.39 | >.00001 | |
| Age | 1.00 | 0.99 | 1.01 | 1.00 | |
| Breast, N = 26,025 | Stage | 0.64 | 0.41 | 1.02 | .0783 |
| Rituximab | 0.90 | 0.59 | 1.38 | 1.00 | |
| Race | 1.60 | 0.82 | 3.13 | 1.00 | |
| Surgery | 1.02 | 0.67 | 1.56 | 1.00 | |
| Extranodal | 1.00 | 0.65 | 1.54 | 1.00 | |
| Radiation | 1.22 | 0.81 | 1.85 | 1.00 | |
| Age | 1.00 | 0.99 | 1.01 | 1.00 | |
| Female genital, N = 11,934 | Stage | 1.13 | 0.60 | 2.15 | 1.00 |
| Rituximab | 1.09 | 0.57 | 2.10 | 1.00 | |
| Race | 1.03 | 0.44 | 2.40 | 1.00 | |
| Surgery | 1.13 | 0.57 | 2.24 | 1.00 | |
| Extranodal | 1.08 | 0.57 | 2.08 | 1.00 | |
| Radiation | 1.01 | 0.52 | 1.97 | 1.00 | |
| Age | 0.99 | 0.97 | 1.00 | .621 | |
| Urinary, N = 26,025 | Stage | 0.86 | 0.55 | 1.35 | 1.00 |
| Rituximab | 1.22 | 0.80 | 1.85 | 1.00 | |
| Race | 2.85 | 1.12 | 7.30 | .00945 | |
| Surgery | 1.29 | 0.82 | 2.03 | 1.00 | |
| Extranodal | 1.17 | 0.75 | 1.81 | 1.00 | |
| Radiation | 0.93 | 0.58 | 1.47 | 1.00 | |
| Sex | 3.54 | 2.10 | 5.95 | >.00001 | |
| Age | 1.02 | 1.01 | 1.03 | >.0000001 | |
| Male genital, N = 14,091 | Stage | 0.78 | 0.53 | 1.15 | 1.00 |
| Rituximab | 0.81 | 0.55 | 1.18 | 1.00 | |
| Race | 1.35 | 0.77 | 2.39 | 1.00 | |
| Surgery | 1.12 | 0.76 | 1.64 | 1.00 | |
| Extranodal | 1.34 | 0.93 | 1.94 | .54 | |
| Radiation | 1.04 | 0.72 | 1.49 | 1.00 | |
| Age | 1.03 | 1.02 | 1.04 | >.00001 | |
| Soft tissue, N = 26,025 | Stage | 0.92 | 0.24 | 3.54 | 1.00 |
| Rituximab | 0.94 | 0.25 | 3.56 | 1.00 | |
| Surgery | 0.61 | 0.15 | 2.43 | 1.00 | |
| Extranodal | 1.38 | 0.38 | 5.21 | 1.00 | |
| Sex | 1.97 | 0.99 | 1.05 | 1.00 | |
| Hematologic, N = 26,025 | Stage | 1.11 | 0.69 | 1.78 | 1.00 |
| Rituximab | 1.99 | 1.26 | 3.14 | >.00001 | |
| Race | 1.25 | 0.65 | 2.41 | 1.00 | |
| Surgery | 1.18 | 0.73 | 1.89 | 1.00 | |
| Extranodal | 0.92 | 0.56 | 1.50 | 1.00 | |
| Radiation | 0.89 | 0.54 | 1.48 | 1.00 | |
| Sex | 1.37 | 0.88 | 2.13 | 1.00 | |
| Age | 1.00 | 0.99 | 1.02 | 1.00 | |
| Unknown, N = 26,025 | Stage | 1.10 | 0.42 | 2.89 | 1.00 |
| Rituximab | 0.66 | 0.27 | 1.59 | 1.00 | |
| Surgery | 1.32 | 0.55 | 3.16 | 1.00 | |
| Extranodal | 0.74 | 0.25 | 2.14 | 1.00 | |
| Radiation | 0.87 | 0.33 | 2.28 | 1.00 | |
| Sex | 1.32 | 0.55 | 3.20 | 1.00 | |
| Age | 1.02 | 1.00 | 1.04 | .594 | |
Abbreviations: HR, hazard ratio; SPM, secondary primary malignancy.
P values and 95% CIs were corrected for multiple comparisons using Bonferroni correction.
The sample size used in each analysis is indicated in the SPM Site column. Sample sizes were too small to assess mediastinum and skeletal SPM sites.
Risk Factors for SPM Development in Exploratory 5-Year Interval Analysis
In an exploratory analysis, we evaluated risk factors for SPM development over 5-year intervals after diagnosis of DLBCL for all SPM locations in aggregate as well as for individual SPM sites, as presented in Supporting Table 2. The following associations were identified, as shown in Figure 3: patients with ES disease were more likely to develop SPM in the period from 0 to 5 years after diagnosis, there was no difference in SPM risk by stage in the period 5 to 10 years after diagnosis, patients with AS disease were more likely to develop SPM in the period 10 to 15 years after diagnosis, and there was no difference in SPM risk by stage beyond 15 years after diagnosis. Furthermore, we found that patients in the ES cohort were more likely to develop colorectal, pancreas, breast, and male genital SPM in the period from 0 to 5 years after diagnosis, and those in the AS cohort were more likely to develop hematologic SPMs in the period 5 to 15 years after diagnosis. Patients with extranodal disease were more likely to develop SPMs in the 0-year to 5-year period, specifically SPM at the foregut and midgut, colorectal, thyroid and thymus, and male genital sites. Patients diagnosed in the postrituximab era were more likely to develop SPM in the 0-year to 5-year period but were less likely to develop SPM in the 5-year to 10-year and 10-year to 15-year periods. Patients diagnosed in the postrituximab era were significantly more likely to develop lymph node, colorectal, hepatobiliary, pulmonary, thyroid and thymus, skin, urinary, and hematologic SPMs in the 0-year to 5-year period.
Figure 3.
Hazard ratios and 95% CIs for secondary primary malignancy (SPM) development (x-axis) between stage of diffuse large B-cell lymphoma at diagnosis (advanced vs early) stratified by 5-year time intervals (y-axis) are illustrated. The vertical dashed line indicates no difference in the hazard between early and advanced stages. Values <1 represent an increased risk of SPM in patients with early stage disease, whereas values >1 represent an increased risk for those with advanced-stage disease.
Survival and Cause of Death
In the Cox proportional hazards model, the interaction among SPM status, stage at diagnosis, and diagnosis in the prerituximab versus postrituximab era was assessed for an impact on survival and found to be significant (P < .0001), as shown in Table 4. The development of an SPM was found to significantly increase the risk of death regardless of stage at diagnosis or rituximab era, with the greatest risk of death seen in patients who had ES disease with SPMs in the postrituximab era (HR, 3.36; 95% CI, 3.02–3.75). However, diagnosis in the postrituximab area was nevertheless associated with overall improved survival compared with diagnosis in the prerituximab era despite increased risk of specific SPMs (see Supporting Figs. 1–3).
TABLE 4.
Cox Proportional Hazards Model Results Assessing Differences in Survival Based on Whether Patients Developed a Secondary Primary Malignancy, Their Stage at Diffuse Large B-Cell Lymphoma DLBCL Diagnosis, and Whether They Were Diagnosed in the Prerituximab or Postrituximab Era, While Adjusting for Age, Sex, and Racea
| Variable | HR | 95% CI | P | |
|---|---|---|---|---|
| Lower | Upper | |||
| Stage | ||||
| Advanced-stage vs early-stage among those with an SPM in the postrituximab era | 1.121 | 0.966 | 1.302 | .1335 |
| Advanced-stage vs early-stage among those with an SPM in the prerituximab era | 1.193 | 1.071 | 1.329 | .0014 |
| Advanced-stage vs early-stage among those without an SPM in the postrituximab era | 1.603 | 1.511 | 1.701 | <.0001 |
| Advanced-stage vs early-stage among those without an SPM in the prerituximab era | 1.553 | 1.492 | 1.617 | <.0001 |
| SPM | ||||
| SPM vs no SPM among those with advanced-stage DLBCL at diagnosis and in the postrituximab era | 2.353 | 2.089 | 2.65 | <.0001 |
| SPM vs no SPM among those with advanced-stage DLBCL at diagnosis in the prerituximab era | 2.005 | 1.825 | 2.202 | <.0001 |
| SPM vs no SPM among those with early-stage DLBCL at diagnosis and in the postrituximab era | 3.364 | 3.015 | 3.754 | <.0001 |
| SPM vs no SPM among those with early-stage DLBCL at diagnosis in the prerituximab era | 2.611 | 2.43 | 2.805 | <.0001 |
| Rituximab | ||||
| Postrituximab era vs prerituximab era among those with an SPM and advanced-stage DLBCL at diagnosis | 0.68 | 0.59 | 0.784 | <.0001 |
| Postrituximab era vs prerituximab era among those with an SPM and early-stage DLBCL at diagnosis | 0.723 | 0.641 | 0.816 | <.0001 |
| Postrituximab era vs prerituximab era among those without an SPM and advanced-stage DLBCL at diagnosis | 0.579 | 0.551 | 0.609 | <.0001 |
| Postrituximab era vs prerituximab era among those without an SPM and early-stage DLBCL at diagnosis | 0.561 | 0.533 | 0.591 | <.0001 |
Abbreviations: DLBCL, diffuse large B-cell lymphoma; HR, hazard ratio; SPM, secondary primary malignancy.
Hazard ratios, 95% CIs, and P values are presented for each level of the interaction. In total, 26,038 patients were included in the analysis, which resulted in 29,381 records. Note that the number of records used in the analysis is larger than the overall number of patients because SPM was treated as a time-dependent covariate, and multiple records were accordingly created for those that developed an SPM to account for the time dependence.
DISCUSSION
In this largest population-based study of SPM in patients with DLBCL to date, we demonstrated first-ever reported differences in the temporal development of SPMs at specific organ sites between patients with ES and AS disease. We found that patients with ES DLBCL were more likely to develop colorectal, pancreas, breast, and male genital SPMs in the period 0 to 5 years after diagnosis, and those with AS DLBCL were more likely to develop hematologic SPMs in the period 5 to 15 years after diagnosis. Our findings highlight that, as patients with DLBCL live longer, their risk of SPM development changes based on their stage at diagnosis and time elapsed since diagnosis.
The association of SPM risk with stage at diagnosis has not been well examined for DLBCL. Okines et al found that AS DLBCL at diagnosis may be a risk factor for SPM, although the odds ratio did not reach statistical significance.17 Tao et al found no significant difference in SPM risk by stage at DLBCL diagnosis.5
Historically, hematologists consider that the risk of SPM increases 5 years after a diagnosis of DLBCL, although recent studies have suggested that patients with DLBCL may have an increased risk of SPM even in the first 5 years after diagnosis.18,19 We found that patients with ES disease were more likely to develop SPM in the period 0 to 5 years after diagnosis, and those with AS disease were more likely to develop SPM in the period 10 to 15 years after diagnosis, as depicted in Figure 3. The most common SPM subtypes in this study were consistent with previous studies of SPMs in NHL.18,20,21
Hematologic SPM
Hematologic SPM was more likely in patients with AS DLBCL during the periods 5 to 10 years and 10 to 15 years after diagnosis. The most common histologies of hematologic SPM were acute myeloid leukemia (27%), myelodysplastic syndrome (21%), plasma cell myeloma (10%), therapy-related myeloid neoplasm (9%), and chronic lymphocytic leukemia/small lymphocytic leukemia (7%). This is consistent with previously published reports of hematologic SPMs after DLBCL, of which acute myeloid leukemia is the most common.2,17–20,22 Our study also confirms previous findings that hematologic SPMs were more likely in the postrituximab era.5,23 Our study did not demonstrate any difference in the risk of lymph node SPM development, depending on stage at diagnosis, with previous studies identifying lymphoma, specifically Hodgkin lymphoma, as a common secondary malignancy after DLBCL.5,19,20,22,24
Male genital SPM
Male genital SPM, which was overwhelmingly prostate adenocarcinoma in the study, was more likely in patients with ES disease and those with extranodal disease in the period 0 to 5 years after diagnosis. Prostate cancer is also a documented SPM after DLBCL.20,21
Colorectal SPM
Colorectal SPM, which was mostly adenocarcinoma by histology in this study, was more likely in patients with ES DLBCL in the period 0 to 5 years after diagnosis and also was more likely in patients diagnosed in the postrituximab era and with extranodal disease in the same period. Colorectal cancer is a well documented SPM after DLBCL diagnosis, although the current study identified additional risk factors.17–22
Breast SPM
Breast SPM, which was mostly carcinoma by histology in this study, was more likely in patients with ES DLBCL and older patients in the period 0 to 5 years after diagnosis. Of note, patients who received radiation were significantly more likely to develop breast SPM in the same period after diagnosis. Breast cancer is a known SPM after DLBCL, with increased risk in younger women.20–22 It has been demonstrated that patients with NHL who receive radiation have an increased risk of breast SPM in prior SEER22 and UK database17 studies. Our study suggests that there is an increased risk of breast SPM as early as 0 to 5 years after DLBCL diagnosis. Verification of SPM development within the irradiated volume was not possible through this database.
Pancreas SPM
Pancreas SPM was more likely in patients with ES DLBCL the period 0 to 5 years after diagnosis. Pancreas SPM has been previously described after an NHL diagnosis,21 although our study suggests as increased risk as early as 0 to 5 years after diagnosis.
Our study identifies novel risk factors for SPM development after DLBCL, including extranodal disease and diagnosis in the postrituximab era after 2001.
In the primary analysis (1973–2010), the presence of extranodal disease was significantly associated with an increased risk of SPM development regardless of stage at diagnosis (Tables 2 and 3), with a significantly higher cumulative incidence of SPM in patients with extranodal disease (Fig. 2). Previous literature has demonstrated that patients with ES DLBCL are more likely to have extranodal disease,25 and our current study adds that an increased risk of SPM risk in patients with ES disease and in those with extranodal disease is seen in the period 0 to 5 years after diagnosis, but not after 5 years.
The primary analysis (1973–2010) also demonstrated an increased risk of SPM development in the postrituximab era after 2001, regardless of stage at diagnosis (Tables 2 and 3). Although a previous meta-analysis has suggested no SPM predisposition among patients with NHL who were exposed to rituximab,26 other studies have demonstrated an increased risk of leukemia, thyroid, pulmonary, hepatobiliary, and melanoma SPM and a decreased risk of Hodgkin lymphoma, colorectal, and breast SPM in the postrituximab era.5,23 This study further elucidates the timeline of SPM risk in the postrituximab era, which was highest in the period 0 to 5 years after diagnosis and became less likely after 5 years (see Supporting Table 2). Previous literature has theorized that rituximab results in B-cell depletion, which reduces antigen presentation to T cells and decreases T-cell–mediated cancer immunosurveillance, thus increasing the risk of future SPM development.10
Our current study adds to the body of literature that ES DLBCL may have a distinct biology compared with AS DLBCL. Roberts et al demonstrated that ES DLBCL was significantly more likely to have germinal center origin on gene expression profiling compared with AS disease,8 and Stephens et al suggested that increased late relapses in ES compared with AS disease may be caused by biological differences between the 2 stages.10 It is plausible that the genetic basis of oncogenesis in ES DLBCL predisposes to distinct SPMs, and that these SPMs may share common genetic origins with ES DLBCL. Furthermore, patients with ES DLBCL are more likely to have extranodal disease (Table 1), and our study further demonstrates that extranodal disease significantly increased the risk of SPM development regardless of stage. Given that all the SPMs with an increased risk in the period 0 to 5 years after a diagnosis of ES DLBCL (colorectal, pancreas, breast, and male genital SPMs) were solid malignancies, it is possible that extranodal disease in ES DLBCL creates a unique genetic milieu that predisposes to the development of secondary solid malignancies within 5 years of diagnosis. This is in contrast to AS DLBCL, which our study demonstrates predisposes patients to hematologic SPMs >5 years after diagnosis, again suggesting a unique genetic profile. Further study of the genetic associations between DLBCL and associated SPMs is warranted.
Overall survival has been extensively studied in DLBCL, with known worse survival in patients with AS DLBCL27,28 and improved survival in the postrituximab era,29 which our study re-demonstrates in Table 3 and Supporting Figures 1 through 3.
However, our study presents the novel finding that SPM development significantly increases the risk of death regardless of stage or rituximab era, with the highest risk of death for patients who have ES DLBCL with an SPM in the postrituximab era (HR, 3.36). Our finding that patients with ES disease, who have longer survival than those with AS disease, have worse survival upon SPM development (Table 4) is consistent with previous research that prolonged survival time predisposes to SPM.30,31 Although poorer survival with SPM development may seem intuitive, the development of SPM after other tumor types has not always demonstrated worse survival. For example, 1 SEER database study found that patients who had both renal cell carcinoma and an SPM had significantly longer survival than patients who had renal cell carcinoma alone.32
The magnitude of risk for SPM, with a 30-year cumulative incidence of 21% for patients with ES DLBCL, and the increased risk for death in patients with SPM point to the importance of clinical vigilance and consideration of SPM screening for this population. Although there are established and effective screening tests for colorectal and breast cancer, there are none specifically for pancreas cancer or hematologic malignancies. Colorectal screening with colonoscopy is not recommended until age 50 years in the general population, and then is recommended every 10 years if the results are negative. It is unknown whether colonoscopy would be of value for treated DLBCL patients if they are aged <50 years or whether a colonoscopy 5 years after DLBCL diagnosis would be worthwhile in patients with ES DLBCL when their risk is the greatest for this SPM subtype. In contrast, the value of prostate-specific antigen screening in the general population is of questionable value with regard to improved cancer mortality.33 It is possible that prostate-specific antigen screening would have greater utility in the high-risk DLBCL population, specifically in patients who have ES disease, in the first 5 years after treatment. The American Society of Hematology recommends avoidance of surveillance computed tomography scans in the patient with treated, asymptomatic DLBCL who has completed therapy.34,35 Identifying subgroups of patients with DLBCL by stage and time since diagnosis for which the incidence of SPM development exceeds the risk of DLBCL recurrence may point to a new indication for imaging in some patients. This requires further study and would be most beneficial in the highest risk DLBCL subsets.
The primary limitation of this study is use of the SEER database, a registry-based data set that does not include several clinical and biological variables, including specific chemotherapy regimens, which may affect SPM development because of unmeasured confounders. For example, this study demonstrated that radiation therapy was not associated with a significant increase in SPM risk; however, because the SEER database codes radiation as “no/unknown” together, it is possible that some of the patients marked with “unknown” radiation status received radiation that was not captured in our analysis, presenting a possible confounder. In addition, because the SEER database does not include information to quantify patient performance status, there may be patient morbidity from SPM development unrelated to mortality. Disease relapse or recurrence is not collected by the SEER database, so it is not possible to account for relapsed DLBCL as a confounder. Although we corrected for multiple comparisons in the primary analyses, no correction was done in the 5-year interval analysis because it was conducted as an exploratory analysis. Finally, because of the observational nature of the study, there may be confounders that resulted in slightly longer follow-up for patients who had ES DLBCL compared with those who had AS DLBCL. Despite the aforementioned shortcomings, our study has shown several novel findings that warrant consideration in future survivorship-focused prospective trials among patients with DLBCL.
In summary, our study presents a novel finding that the risk of SPM development in patients diagnosed with DLBCL is not uniform for all patients and changes over time based on several risk factors, most importantly stage at diagnosis and time since diagnosis. We also observed that development of SPM increased the risk of death regardless of DLBCL stage at diagnosis. The strength of this study is the analysis of a large, population-based cohort that permits robust detection of statistical differences in SPM development between patients with ES and AS DLBCL. In addition, this study analyzes SPM risk over the entire study period and at 5-year intervals, permitting a more nuanced description of SPM risk over time. In conclusion, for patients with recently diagnosed DLBCL in the current postrituximab era whose risk of SPM development is highest in the first 5 years after diagnosis, it may be prudent for survivors to undergo site-specific surveillance screening based on the period of time since diagnosis. Concerted collaborative efforts to reevaluate current survivorship guidelines are warranted.
Supplementary Material
FUNDING SUPPORT
No specific funding was disclosed.
Footnotes
Preliminary data from this study were presented in part at the 2018 American Society of Clinical Oncology (ASCO) Annual Meeting; June 1–5, 2018; Chicago, Illinois.
CONFLICT OF INTEREST DISCLOSURES
Manali Kamdar reports personal fees from Seattle Genetics, Celgene, Pharmacyclics, and AstraZeneca outside the submitted work. The remaining authors made no disclosures.
REFERENCES
- 1.Donin N, Filson C, Drakaki A, et al. Risk of second primary malignancies among cancer survivors in the United States, 1992 through 2008. Cancer. 2016;122:3075–3086. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Morton LM, Curtis RE, Linet MS, et al. Second malignancy risks after non-Hodgkin’s lymphoma and chronic lymphocytic leukemia: differences by lymphoma subtype. J Clin Oncol. 2010;28:4935–4944. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Miller KD, Siegel RL, Lin CC, et al. Cancer treatment and survivorship statistics. CA Cancer J Clin. 2016;66:271–289. [DOI] [PubMed] [Google Scholar]
- 4.Pirani M, Marcheselli R, Marcheselli L, et al. Risk for second malignancies in non-Hodgkin’s lymphoma survivors: a meta-analysis. Anna Oncol. 2011;22:1845–1858. [DOI] [PubMed] [Google Scholar]
- 5.Tao L, Clarke CA, Rosenberg AS, et al. Subsequent primary malignancies after diffuse large B-cell lymphoma in the modern treatment era. Br J Haematol. 2017;178:72–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Cho SF, Wu WH, Yang YH, et al. Risk of second primary cancer in patients with B-cell non-Hodgkin lymphoma receiving rituximab-containing chemotherapy: a nationwide population-based study. Anticancer Res. 2015;35:1809–1814. [PubMed] [Google Scholar]
- 7.Friedberg JW, Fisher RI. Diffuse large B-cell lymphoma. Hematol Oncol Clin North Am. 2008;22:941–952. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Roberts RA, Rimsza LM, Staudt L, et al. Gene expression differences between low and high stage diffuse large B cell lymphoma (DLBCL). Blood. 2006;108:809. [Google Scholar]
- 9.Larouche JF, Berger F, Chassagne-Clement C, et al. Lymphoma recurrence 5 years or later following diffuse large B-cell lymphoma: clinical characteristics and outcome. J Clin Oncol. 2010;28:2094–2100. [DOI] [PubMed] [Google Scholar]
- 10.Stephens DM, Li H, LeBlanc ML, et al. Continued risk of relapse independent of treatment modality in limited-stage diffuse large B-cell lymphoma: final and long-term analysis of Southwest Oncology Group Study S8736. J Clin Oncol. 2016;34:2997–3004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Vogt A, Schmid S, Heinimann K, et al. Multiple primary tumours: challenges and approaches, a review. ESMO Open. 2017;2:e000172. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.R Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing; 2015. [Google Scholar]
- 13.SAS Institute, Inc. SAS 9.3 Language Reference. SAS Institute, Inc; 2012. [Google Scholar]
- 14.Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. 1999;94:496–509. [Google Scholar]
- 15.Gerds TA, Scheike TH, Blanche P, Ozenne B. riskRegression: Risk Regression Models and Prediction Scores for Survival Analysis with Competing Risks. R package version 1.4.3. Accessed November 26, 2018. https://cran.r-project.org/web/packages/riskRegression/ [Google Scholar]
- 16.Gray B Cmprsk: Subdistribution Analysis of Competing Risks. R package version 2.2–7. Accessed November 26, 2018. https://cran.r-project.org/web/packages/cmprsk/index.html [Google Scholar]
- 17.Okines A, Thomson CS, Radstone CR, et al. Second primary malignancies after treatment for malignant lymphoma. Br J Cancer. 2005;93:418–424. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Ghimire KB, Shah BK. Second primary malignancy in diffuse large B cell lymphoma (DLBCL). Blood. 2014;124:2606. [Google Scholar]
- 19.Gangavarapu LS, Surapaneni R. Diffuse large B-cell lymphoma survivorship: risk of second malignancies in long term survivors. Blood. 2015;126:3921. [Google Scholar]
- 20.Wang Y, Wang Y, Thanarajasingam G, et al. Second primary malignancies in patients with non-Hodgkin lymphoma. Blood. 2017;130(suppl 1):2152. [Google Scholar]
- 21.Bermejo JL, Pukkala E, Johannesen TB, et al. Age-time risk patterns of solid cancers in 60 901 non-Hodgkin lymphoma survivors from Finland, Norway and Sweden. Br J Haematol. 2014;164:675–683. [DOI] [PubMed] [Google Scholar]
- 22.Tward JD, Wendland MM, Shrieve DC, et al. The risk of secondary malignancies over 30 years after the treatment of non-Hodgkin lymphoma. Cancer. 2006;107:108–115. [DOI] [PubMed] [Google Scholar]
- 23.Keegan THM, Clarke CA, Rosenberg AS, et al. Subsequent primary malignancies (SPMs) after diffuse large B-cell lymphoma (DLBCL) in the modern treatment era [abstract]. J Clin Oncol. 2016;34(suppl 1):7543. [Google Scholar]
- 24.Moser EC, Noordijk EM, van Leeuwen FE, et al. Risk of second cancer after treatment of aggressive non-Hodgkin’s lymphoma; an EORTC cohort study. Haematologica. 2006;91:1481–1488. [PubMed] [Google Scholar]
- 25.Castillo JJ, Winer ES, Olszewski AJ. Sites of extranodal involvement are prognostic in patients with diffuse large B-cell lymphoma in the rituximab era: an analysis of the Surveillance, Epidemiology and End Results database. Am J Hematol. 2014;89:310–314. [DOI] [PubMed] [Google Scholar]
- 26.Fleury I, Chevret S, Pfreundschuh M, et al. Rituximab and risk of second primary malignancies in patients with non-Hodgkin lymphoma: a systematic review and meta-analysis. Ann Oncol. 2016;27:390–397. [DOI] [PubMed] [Google Scholar]
- 27.Sehn LH, Berry B, Chhanabhai M, et al. The revised International Prognostic Index (R-IPI) is a better predictor of outcome than the standard IPI for patients with diffuse large B-cell lymphoma treated with R-CHOP. Blood. 2007;109:1857–1861. [DOI] [PubMed] [Google Scholar]
- 28.Sehn LH, Gascoyne RD. Diffuse large B-cell lymphoma: optimizing outcome in the context of clinical and biologic heterogeneity. Blood. 2015;125:22–32. [DOI] [PubMed] [Google Scholar]
- 29.Shah BK, Bista A, Shafii B. Survival in advanced diffuse large B-cell lymphoma in pre- and post-rituximab eras in the United States. Anticancer Res. 2014;34:5117–5120. [PubMed] [Google Scholar]
- 30.Shan S, She J, Xue ZQ, et al. Clinical characteristics and survival of lung cancer patients associated with multiple primary malignancies. PLoS One. 2017;12:e0185485. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Lv M, Zhang X, Shen Y, et al. Clinical analysis and prognosis of synchronous and metachronous multiple primary malignant tumors. Medicine (Baltimore). 2017;96:e6799. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Chakraborty S, Tarantolo SR, Batra SK, et al. Incidence and prognostic significance of second primary cancers in renal cell carcinoma. Am J Clin Oncol. 2013;36:132–142. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Pinsky PF, Prorok PC, Yu K, et al. Extended mortality results for prostate cancer screening in the PLCO trial with median follow-up of 15 years. Cancer. 2017;123:592–599. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.American Society of Hematology. Limit surveillance computed tomography (CT) scans in asymptomatic patients following curative-intent treatment for aggressive lymphoma. Accessed November 26, 2018. http://www.choosingwisely.org/clinician-lists/american-society-hemaology-limit-surveillance-ct-scans-following-treatment-for-lymphoma/
- 35.Lin TL, Kuo MC, Shih LY, et al. Value of surveillance computed tomography in the follow-up of diffuse large B-cell and follicular lymphomas. Ann Hematol. 2012;91:1741–1745. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.