Diffuse Large B-Cell Lymphoma in Chile: The Impact of Combined CHOP Plus Rituximab in the Public Health System

PURPOSE

Diffuse large B-cell lymphoma (DLBCL) is the most common lymphoma subtype. The purpose of this study was to evaluate the clinical features, prognostic factors, and results of DLBCL that was treated in the cancer centers of the public health system in Chile and compare cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) with rituximab with cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP).

PATIENTS AND METHODS

Patients age > 15 years who were treated in 18 cancer centers in the country between 2001 and 2017 were included. The Kaplan-Meier method was used to calculate overall survival (OS), and Cox proportional hazard regression modeling was used to evaluate the effect of the addition of rituximab to CHOP on OS.

RESULTS

A total of 1,807 patients were evaluated. The median age at diagnosis was 62 (range, 15-95) years, with a female predominance (53%). Half of the patients were age ≥ 60 years. Serology for HIV infection was positive in 5% of cases (96 cases). International Prognostic Index scores were available for 90% of patients, of which 45% had low-risk, 25% low-intermediate-risk, 18% high-intermediate-risk, and 11% high-risk scores. CHOP was administered to 986 patients (55%; median follow-up, 13.2 years) and R-CHOP to 821 patients (45%; median follow-up, 8.4 years). R-CHOP was associated with superior OS compared with CHOP (5-year 66% v 48%, and 10-year 53% v 35%; P < .001).

CONCLUSION

Rituximab improved the survival of patients with DLBCL diagnosed and treated in Chile. The benefit was sustained over time, with curative rates of > 50%. This intervention shows that the inclusion of this biological drug justified the expenses incurred by the Ministry of Health in the National Lymphoma Protocols in Chile.

INTRODUCTION

Diffuse large B-cell lymphoma (DLBCL) is the most common non-Hodgkin lymphoma, accounting for 30%-40% of all cases worldwide^1,2 as well as in Chile (38%).³ The clinical course of DLBCL is aggressive, although it is potentially curable. Approximately 50% of patients are cured, 10%-15% are refractory, and 20%-30% relapse. The standard treatment is rituximab with cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP), since it demonstrated a significant improvement in survival, compared with CHOP, in patients age older than 60 years^4,5 and in patients age younger than 60 years.⁶ The response persists over time, with a 5-year progression-free survival of approximately 80% in early-stage disease and more than 50% in advanced disease.^7-9

CONTEXT

• Key Objective

• How do patients with diffuse large B-cell lymphoma (DLBCL) from Chile, a middle-income Latin American country, respond to chemoimmunotherapy, which was largely evaluated in clinical trials in developed countries? Did the survival of patients with DLBCL in Chile improve with the addition of anti-CD20 antibody, as observed in such clinical trials?

• Knowledge Generated

• A large cohort of 1,807 patients diagnosed and treated in Chile between 2001 and 2017 were evaluated. The addition of rituximab to chemotherapy improved the overall survival (OS) in patients with DLBCL compared with chemotherapy alone with 5-year OS rates of 66% versus 48%, respectively, and 10-year OS rates of 53% versus 35%, respectively.

• Relevance

• Rituximab with cyclophosphamide, doxorubicin, vincristine, and prednisone is the standard frontline therapy for DLBCL, and the expenses incurred by the public health system was justified in Chile and likely also in other developing countries.

The Adult Cancer National Program (PANDA) of the Ministry of Health of Chile has financed lymphoma treatment since 1988, and rituximab has been associated with chemotherapy since 2007. At the same time, in 2005, lymphoma was incorporated into the System of Explicit Guaranties of Health (GES), which guaranteed access, opportunity, and financial coverage to patients in the public as well as the private Chilean health system.

The aim of this study was to analyze the demographic and clinical characteristics as well as the results of patients with DLBCL who had been treated and registered in the cancer centers of the public health system of Chile, since immunohistochemical studies were performed routinely to establish the B- or T-cell origin.

PATIENTS AND METHODS

Patients

Patients age > 15 years old were analyzed. They were diagnosed and treated at 18 cancer centers in Chile, between 2001 and 2017. They were prospectively registered in a database at the Cancer Unit of the Ministry of Health. In 2001, immunohistochemical testing was routinely performed at all centers for the diagnosis of lymphoma. This study was approved by the Scientific Ethics Committee of Servicio de Salud Metropolitano Oriente in Santiago. All patients signed an informed consent form for treatment, and their data were collected before treatment.

Histopathologic diagnosis was centralized and performed by a reference pathologist (V.M.C.). Staging followed the National Guidelines updated according to the literature. It included a complete blood count, biochemistry tests including lactic dehydrogenase, computed tomography (CT) scan, bone marrow biopsy, and viral serology (HIV, hepatitis B, hepatitis C, and human T-lymphotropic virus type 1). HIV-positive patients were included in this study. Ann Arbor clinical staging was used from stages I-IV. Patients were stratified according to the standard International Prognostic Index (IPI).¹⁰ They were also stratified according to the revised IPI score.¹¹ Patients with primary CNS lymphoma, transformed indolent lymphoma, or those who received chemotherapy regimens other than CHOP or R-CHOP were excluded.

Treatment

For early stages (I/II), four cycles of CHOP between 2001 and 2006 and R-CHOP from 2007 to 2017 plus 30 Gy of involved-field radiotherapy (RT) were administered. For advanced stages (III/IV), six to eight cycles of CHOP and R-CHOP were administered, depending on the period and treating physician's discretion. Rituximab was progressively introduced throughout the country according to the availability of the drug, not on the basis of patient's stage or severity of disease.

Response criteria were defined according to the International Consensus for Lymphoma.¹² Complete remission (CR) was defined as no evidence of disease; partial remission was defined as ≥ 50% reduction of masses; stable disease was defined as ≤ 50% reduction of masses; and relapse (after CR)/progressive disease (partial remission/stable disease) was defined as new lesions or an increase of ≥ 50% of masses. The response was evaluated using CT scans after three cycles of therapy (interim) and after therapy completion. Positron emission tomography-CT scanning was not routinely used for staging and response assessment in Chile until its introduction in 2017.

Follow-Up

Patients underwent clinical evaluation every 4 months in the first year from diagnosis, every 6 months from the second to the fifth year, and annually thereafter. A CT scan was performed as clinically indicated. Refractory or relapsed cases were mainly treated with etoposide, cisplatin, methylprednisolone, and cytarabine, etoposide, prednisone, vincristine, cyclophosphamide, and doxorubicin, ifosfamide, carboplatin, and etoposide, rituximab, cyclophosphamide, vincristine, and prednisone, or palliative care. Autologous stem-cell transplantation was performed during the second CR for patients age < 40 years. The cause of death was obtained from Civil Registry Death Certificates. This is filled by the physician who assists the patient or the one who has seen him or her. Then, the Civil Registry sends the database to the Department of Statistic and Health Information (DEIS: Departamento de Estadisticas e Información en Salud) of the Ministry of Health daily.

Statistical Analysis

Clinical characteristics and response to treatment were compared between those treated in the era before and after rituximab if they received at least one cycle of chemotherapy. Patient characteristics are presented using descriptive statistics. Overall survival (OS) was defined as the time from the date of diagnosis to death or the last follow-up. Progression-free survival could not be calculated, as data were not available for two thirds of the cases. OS was estimated using the Kaplan-Meier method and compared using the log-rank test. Statistical calculations and graphs were generated using STATA version 17 (StataCorp, College Station, TX). P values < .05 were considered statistically significant.

RESULTS

Patients' Characteristics

Between 2001 and 2017, 1,883 patients with DLBCL were diagnosed. We included 1,807 cases in this analysis; 76 cases were excluded because they received regimens other than CHOP or R-CHOP. Participating centers and the number of cases included Hospital del Salvador (n = 509), Instituto Nacional del Cancer (n = 210), Barros Luco-Trudeaux (n = 167), Sótero del Río (n = 165), San Juan de Dios (n = 146), San Borja Arriaran (n = 110), Valdivia (n = 133), Gustavo Fricke (n = 85), Talca (n = 85), San José (n = 74), Osorno (n = 44), La Serena (n = 38), Van Buren (n = 33), Arica (n = 38), Las Higueras (n = 3), Antofagasta (n = 19), Concepción (n = 7), and Temuco (n = 6).

Table 1 presents the baseline patient characteristics. The median age at diagnosis was 62 (range, 15-95) years, and 56% (1,010 cases) of patients were age ≥ 60 years, with a higher proportion of patients age ≥ 60 years in the CHOP group (P < .01). Ninety-six patients (5%) had a positive serology for HIV, with a higher proportion in the CHOP group (P < .01). The clinical stage distribution was as follows: stage I, 120 cases (6.6%); IE, 241 cases (13.3%); II, 278 cases (15.3%); IIE, 263 cases (14.5%); III, 302 cases (16.7%); and IV, 603 cases (33.4%). Approximately one third (28%) of patients presented with an extranodal site of involvement, stage IE/IIE (504 cases). The GI tract was the most common extranodal site, 211/504 (42%). Among them, 158 cases (75%) were gastric, 45 (21%) were intestinal, 3 (1%) were rectal, and 5 (2%) had multiple sites. Other extranodal sites included Waldeyer ring (n = 75), thyroid (n = 30), oral cavity (n = 29), bone (n = 28), nasal (n = 20), testicle (n = 13), gynecologic (n = 12), breast (n = 11), skin (n = 8), parotid (n = 7), and other less common. IPI and R-IPI were available for 1,619 cases (90%). The IPI and R-IPI distributions are presented in Table 1.

TABLE 1.

Baseline Characteristics of the 1,807 Patients With Diffuse Large B-Cell Lymphoma, Comparing CHOP With R-CHOP

Treatment

CHOP and R-CHOP were administered to 985/1,807 (54.5%) and 821/1,807 (45.5%) patients, respectively. The proportion of patients who received R-CHOP increased over time (Fig 1). Of 904 patients with advanced disease, 525 (58%) received six cycles and the remaining received eight cycles, per the treating physician's discretion. In the whole group, 531/1,807 (29%) patients had RT, and most of them (70%) had a limited stage. Of 903 patients with limited-stage disease, 270 patients (30%) did not receive RT because the patient had surgical resection of affected organs, had a contraindication for radiation, or the treating physician did not considered RT suitable. Of 904 patients with advanced disease, 161 (18%) received RT for bulky disease (> 10 cm in diameter). Of 159 cases with partial/refractory or relapse disease, 125 (79%) received etoposide, cisplatin, methylprednisolone, and cytarabine, and the rest were quite diverse, such as etoposide, prednisone, vincristine, cyclophosphamide, and doxorubicin, ifosfamide, carboplatin, and etoposide, rituximab, cyclophosphamide, vincristine, and prednisone, rituximab, RT, or palliative care. Only two patients underwent autologous stem-cell transplantation for consolidation. The CR rate was significantly higher in the R-CHOP group (n = 572, 70%) than in the CHOP group (n = 587, 59%; P < .001). Twenty-one patients had CNS prophylaxis with intrathecal methotrexate; nine had bone marrow involvement, five skeletal bone involvement, three CNS, three nasal cavity, and one had testicular involvement. Thirteen are alive: 3/7 (43%) in the CHOP group and 10/14 (71%) in the R-CHOP group. There were only two patients with CNS relapse in the CHOP era and none in the R-CHOP era.

FIG 1.

Trends of usage of CHOP and R-CHOP in patients with DLBCL in Chile between 2001 and 2017. CHOP, cyclophosphamide, doxorubicin, vincristine, and prednisone; DLBCL, diffuse large B-cell lymphoma; R-CHOP, rituximab plus CHOP.

Survival Analyses

The median follow-up of the entire group was 10.3 years (95% CI, 10.1 to 10.7), with a median follow-up of 13.2 years (95% CI, 12.8 to 13.7) for the CHOP group and 8.4 years (95% CI, 8 to 8.7) for the R-CHOP group (Fig 2A). The median OS for the entire group was 6.8 years (95% CI, 6.1 to 10.7). The median OS for CHOP was 4.5 years, whereas that for R-CHOP was not reached (log-rank P < .001). The five- and 10-year OS rates for the entire cohort were 56% and 43%, respectively. For those who received CHOP, these rates were 48% and 35%, respectively, whereas the rates of those who received R-CHOP were 66% and 53%, respectively.

FIG 2.

Kaplan-Meier estimates of OS according to (A) therapy received, (B) period of diagnosis and treatment, (C) therapy received and age, (D) IPI score, (E) R-IPI score, (F) and in patients with relapsed DLBCL. CHOP, cyclophosphamide, doxorubicin, vincristine, and prednisone; DLBCL, diffuse large B-cell lymphoma; IPI, International Prognostic Index; OS, overall survival; R-CHOP, rituximab plus CHOP.

The prognosis of patients improved over the 17-year period of this study (Fig 2B), with a median OS of 5 years (95% CI, 3.9 to 6.4) for patients diagnosed between 2001 and 2005, 7.5 years (95% CI, 9.4 to not reached) for patients diagnosed between 2006 and 2011, and not reached for patients treated between 2012 and 2017 (log-rank P = .003). There was no difference in the OS according to sex or HIV status (data not shown).

In the CHOP group, patients age < 60 years had a 5-year OS rate of 56% (95% CI, 51 to 61) and a 10-year OS rate of 48% (95% CI, 42 to 52). Patients age ≥ 60 years had a 5-year OS rate of 43% (95% CI, 39 to 46) and a 10-year OS rate of 26% (95% CI, 23 to 30; log-rank P < .001). In the R-CHOP group, patients age < 60 years had a 5-year OS rate of 74% (95% CI, 69 to 78) and a 10-year OS rate of 67% (95% CI, 61 to 71). In patients age ≥ 60 years, the 5-year OS rate was 58% (95% CI, 52 to 62) and the 10-year OS rate was 39% (95% CI, 33 to 45; log-rank P < .001; Fig 2C).

According to the IPI score, the median OS was 15.6 years (95% CI, 14 to not reached) for low-risk, 6.8 years (95% CI, 5.3 to 8.4) for low-intermediate-risk, 2.2 years (95% CI, 1.3 to 3.0) for high-intermediate-risk, and 1.1 years (95% CI, 0.7 to 1.3) for high-risk patients (log-rank P < .001; Fig 2D). The R-IPI score separated patients into three prognostic groups, with median survival times of not reached for low-risk, 9.1 (95% CI, 7.6 to 10.4) years for intermediate-risk, and 1.3 years (95% CI, 1.1 to 2.2) for high-risk patients (log-rank P < .001; Fig 2E).

We constructed two multivariate Cox proportional hazard regression models to evaluate the effect of R-CHOP on OS. In model 1, R-CHOP (v CHOP) was evaluated after adjusting for age, Eastern Cooperative Oncology Group performance status, lactic dehydrogenase level, clinical stage, extranodal involvement, and year of diagnosis. In this model, which included 1,589 observations, R-CHOP was associated with a 23% lower risk of death than CHOP (hazard ratio, 0.76; 95% CI, 0.70 to 0.83; P < .001; Table 2). In model 2, R-CHOP (v CHOP) was evaluated by adjusting for the standard IPI score and year of diagnosis. In this model, which included 1,619 observations, R-CHOP was associated with a 26% lower risk of death than CHOP (hazard ratio, 0.74; 95% CI, 0.68 to 0.80; P < .001; Table 2).

TABLE 2.

Multivariate Cox Proportional-Hazard Regression Models Evaluating the Effect of R-CHOP Versus CHOP in the Overall Survival of Patients With Diffuse Large B-Cell Lymphoma Diagnosed in Chile Between 2001 and 2017

Among 240 patients with data on relapse, 153 were in the CHOP group and 87 in the R-CHOP group. The median survival time from relapse was 0.5 years (95% CI, 0.4 to 0.7), with a 5-year OS rate of 19% (95% CI, 14 to 24; Fig 2F).

Cause of Death

The cause of death was reported in 1,008 patients who received CHOP and R-CHOP. The most common causes included lymphoma progression (70% and 72%, respectively), toxicity/complications of treatment (8% in both), cardiovascular disease (6% in both), cancer other than lymphoma (5% in both), and infections not related to chemotherapy (6% and 4%, respectively). Other causes, such as respiratory failure, chronic liver damage, GI bleeding, dementia, suicide, or accidents, were less common (Table 3).

TABLE 3.

Causes of Death on Patients With DLBCL Diagnosed in Chile Between 2001 and 2017

The cause of death in 226 patients who died 5 years after diagnosis, in CHOP and R-CHOP groups, was lymphoma progression but with a lower frequency (44% and 43%, respectively) compared with those who died within 5 years of diagnosis. However, neoplasia other than lymphoma was most common 5 years after diagnosis (5%-12% for CHOP and 5%-14% for R-CHOP), as well as cardiovascular disease (6%-13% for CHOP and 6%-11% for R-CHOP; Table 3).

The 52 patients who died of cancers other than lymphoma were evenly distributed in both groups. The most common locations were the GI tract (n = 26), gastric (n = 11), gallbladder (n = 4), pancreas (n = 2), ampulla of Vater (n = 2), oropharynx (n = 2), esophagus (n = 2), colon (n = 2), and liver (n = 1). Other cancers included acute myeloid leukemia (n = 7), prostate (n = 6), gynecologic (n = 3), skin (n = 2), and bladder, testicle, chronic myeloid leukemia, CNS, and lung (n = 1 each). The cancer was of unknown origin in six cases.

The causes of death due to cardiovascular disease include myocardial infarction, heart failure, stroke, or pulmonary thromboembolism. Infection/sepsis outside of the treatment was mainly of pulmonary or urinary origin. In seven cases, the cause of death was not determined.

DISCUSSION

To our knowledge, our results show the evaluation of the largest cohort of patients with DLBCL in Chile treated in the public health system. This demonstrates that the addition of rituximab to CHOP was a successful policy, significantly improving the outcome of DLBCL, with 18% absolute improvement in 5- and 10-year OS with R-CHOP compared with those treated with CHOP alone.

If we compare our results with those in the literature, they are quite similar in patients age ≥ 60 years. For example, the GELA study in patients age ≥ 60 years reported 5-year OS rates of 45% and 58% with CHOP and R-CHOP, respectively, similar to our results of 43% and 58%, respectively. For 10-year OS, the results were slightly better for CHOP and R-CHOP, 27.6% and 43.5%, respectively, whereas ours were 26% and 39%, respectively.^4,8,9 Regarding patients age < 60 years, our results were significantly inferior, since the German group MINTH reported 6-year OS rates of 80% and 90%, respectively, with CHOP and R-CHOP, superior to our results of 56% and 74%, respectively.^6,7 Regardless of this difference, improvement in both age groups compared with previous treatment is impressive, since survival rates improved in the range of 15%-18%.

Demographic characteristics were similar to those described in Europe and the United States, with a median age of 62 years, as previously described.^3,13 Compared with other Latin American countries, it is noteworthy that the age of presentation is higher in Chile, as in developed countries.^14,15 It may well be due to the higher life expectancy at birth¹⁵ in our country, which is superior to the rest of Latin America and similar to that of developed countries belonging to Organization for Economic Co-operation and Development. The higher life expectancy in Chile may be due to several factors, including the high health coverage of the public health system network, higher socioeconomic levels, and better education, with one of the highest literacy rates in Latin America.^16,17 Female sex predominance was striking (53%). Two previous studies from Chile reported a higher proportion of women among patients with extranodal lymphoma and multiple myeloma.^13,18 Almost everywhere, male sex is predominant in DLBCL, except in Brazil, according to the study of the International Non-Hodgkin Lymphoma Classification Project.¹⁴ The reasons behind this finding are not well understood.

Studies in other Latin American countries, such as Uruguay, Mexico, Peru, Brazil, and Argentina,^19-25 showed a lower median age of 56-58 years and an improved outcome with rituximab, except in two Mexican studies.^22,23 It is remarkable that in the study by Candelaria et al²⁰ from Mexico, a federal program financed the treatment with no cost for the patients, in a similar way to ours, both countries with limited resources. It is important to note that since 2017, biosimilars have come up to the pharmaceutical armamentarium with identical efficacy.^26-31

The long follow-up of the entire cohort, 10 years, allowed us to observe that if a patient survives for more than 5 years, the patient is at risk of other neoplasia, besides lymphoma, or cardiovascular disease, similar to the rest of the population of similar age. Our results show that lymphoma is the main cause of death within 5 years of diagnosis. However, in patients surviving more than 5 years from diagnosis, death from other tumors is more prevalent, 12%-14%, which is slightly higher than that reported in the GELA study (5%-9%). Regarding the origin of cancer, gastric neoplasia was the most common, a reflection of the epidemiology of cancer in Chile, since it is one of the most prevalent cancers, with a high mortality rate among men in our country, similar to prostate and lung.^32-34 Gallbladder cancer is the sixth leading cause of mortality in Chile.³⁵ We can say that patients cured from lymphoma are at risk of developing the same neoplasia and unrelated cause of death, compared with the general population. The incidence of cardiovascular disease has also increased.³⁵

The prognosis of patients who are refractory or relapse after R-CHOP therapy is ominous.^36-38 In our case, the 5-year OS rate after relapse was 19%. Some advances in treatment have gone from shortening intervals between cycles,³⁹ consolidating with high doses of chemotherapy and autologous stem-cell transplant^40,41 or obinutuzumab, antibody anti-CD20 type II, compared with rituximab,^42,43 or polatuzumab, selinexor, or CAR T cells.^44-46 Positron emission tomography-CT–guided intensified therapy has also been tried⁴⁷ when this is positive. To date, no other treatment has replaced R-CHOP-21 as a first-line therapy.

To optimize therapy in the future, it will be imperative to understand the molecular heterogeneity of DLBCL and to investigate novel targeted agents.⁴⁸ It is necessary to know the cell of origin, whether there is a double or triple expressor of MYC and BCL2 or double- or triple-hit case, since these are higher-grade lymphomas,⁴⁹ possibly requiring more aggressive therapy. The ability to detect predictive biomarkers is crucial for adapted risk therapies.

The limitations of our study include a lack of relapse date in approximately two thirds of cases and a lack of a centralized histopathology review, although more than 80% of cases were reviewed by one pathologist. An important limitation was that this study was registry-based and therefore, there could have been biases introduced on the basis of the quality of the data registered. Additionally, data were not available on dose intensity or dose delays.

In conclusion, the addition of rituximab to CHOP was associated with significant survival improvement in patients with DLBCL treated in the public health system in Chile. The funding of lymphoma in the GES system was a cost-effective measure on the basis of the results of this study.

AUTHOR CONTRIBUTIONS

Conception and design: María Elena Cabrera, Camila Peña, Jorge J. Castillo

Financial support: All authors

Administrative support: María Elena Cabrera, Camila Peña, Jorge J. Castillo

Provision of study materials or patients: María Elena Cabrera, Camila Peña, Vivianne Lois, Hernán Rojas, Valeska Vega, Alvaro Pizarro, Susana Calderon, Christine Rojas, Augusto Aspillaga, M. Luisa Gonzalez, Marvila Intriago, Bernardita Rojas, Cecilia Hales, Jacqueline Oliva, Mónica Romero, Marisa Capurro

Collection and assembly of data: María Elena Cabrera, Camila Peña, Pilar Leon, Vivianne Lois, Hernán Rojas, Valeska Vega, Alvaro Pizarro, Susana Calderon, Christine Rojas, M. Luisa Gonzalez, Marvila Intriago, Bernardita Rojas, Cecilia Hales, Jacqueline Oliva, Mónica Romero, Marisa Capurro, Jorge J. Castillo

Data analysis and interpretation: María Elena Cabrera, Camila Peña, Pilar Leon, Augusto Aspillaga, Jorge J. Castillo

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/go/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).