nueva página del texto (beta)
Inglés (pdf)
Artículo en XML
Referencias del artículo
Traducción automática
Enviar artículo por email
Citado por SciELO 
Similares en
SciELO 
Permalink
Introduction
Histiocytic disorders (HDs) are diverse diseases derived from myeloid progenitors of monocytic, histiocytic, and dendritic lines. According to the current classification, these disorders can be divided into C, H, L, M, and R groups. Langerhans cell histiocytosis (LCH) belongs to the L group, which also includes Erdheim-Chester disease (ECD), indeterminate cell histiocytosis, and mixed HDs (ECD-LCH)1,2. In Mexico, the incidence of HDs is 4.3 cases per million children under 18 years of age3.
The clinical presentation of LCH is heterogeneous, ranging from single lesions to the involvement of multiple organs, including the liver, spleen, and bone marrow. Similar findings have been noted in different countries4-6.
The Histiocyte Society has observed in its studies that the administration of vinblastine and prednisone confers a survival rate of up to 99% in patients without at-risk organs; the survival rate falls to 84% in patients with liver, spleen, and/or bone marrow involvement7. Other studies have demonstrated the usefulness of other drugs, such as cytarabine, and the importance of prolonged treatments of up to 12 months8,9.
The BRAFV600]E somatic mutation was recently described in up to 57% of patients with LCH10,11. This mutation has been linked to greater involvement in organs at risk (88%), more significant reactivation at 5 years (43% vs. 28%), and increased resistance to chemotherapy (22% vs. 3%)12,13. The disease relapses frequently, so new treatment strategies are constantly being explored based on drugs such as cladribine, clofarabine, bisphosphonates, vemurafenib, and dabrafenib14-21.
Since 2010, treatment for children with LCH in Mexico has been based on the provisions of the LCH III protocol of the Histiocyte Society and the National Technical Protocols of the National Council for the Prevention and Treatment of Cancer in Childhood and Adolescence22.
This study evaluated pediatric patients’ clinical characteristics, treatment, and survival across 19 hospitals accredited for caring for children with cancer in Mexico. This investigation focused on diagnoses between January 2010 and December 2018.
Methods
An observational retrospective study was conducted in hospitals accredited for treating children with cancer. All centers throughout Mexico were invited to participate, of which nineteen hospitals accepted this invitation. All patients included were newly diagnosed with LCH, under 18 years of age, and diagnosed between 2010 and 2018. The patients’ diagnoses were based on histopathology and immunohistochemistry and, in some cases, electron microscopy (6.3%). Extension studies were performed at the patients’ treatment centers. All participants were classified according to a clinical group: (1) Group 1 presented involvement of at-risk organs (liver, spleen, and bone marrow); (2) Group 2 did not exhibit at-risk organ involvement; and (3) Group 3 exhibited involvement of the central nervous system and spinal cord. We collected information regarding initial clinical manifestations, affected sites, laboratory studies at diagnosis, histopathology studies, relapse or progression, survival, and mortality. Since 2010, most institutions in Mexico have followed the national protocol based on the Histiocyte Society’s LCH III described in table 1: Patients in Group 1 are given a 12-month course of chemotherapy, and patients in Groups 2 and 3 receive 6 months of chemotherapy. The patients were followed up at the discretion of each center.
Table 1 Treatment protocol in children with LCH in Mexico
| Induction | Vinblastine 6 mg/m2 IV weekly for 6 or 12 weeks |
| If there is a complete response at the end of the first induction, maintenance is performed; if there is not, another 6 weeks are given. | Prednisone 40 mg/m2 oral for 6-12 weeks |
| Maintenance | Vinblastine 6 mg/m2 IV every 3 weeks |
| Low risk - 25 weeks | |
| High risk - 52 weeks | Prednisone 40 mg/m2 oral for 5 days, every 3 weeks |
Relapse was defined as disease reactivation after a documented complete response, defined as the disappearance of all metabolic activity in the evaluation after treatment through imaging studies.
Since this study was retrospective, our sample was determined by convenience; patients registered at each center during the study period were included.
Patients whose records contained at least 80% of the required information were included in the study.
Statistical analysis
We performed univariate analysis using central-tendency tests to better understand the characteristics of the patient sample and to establish how each variable was distributed. We calculated means and standard deviations for continuous numerical variables with normal distributions. We calculated the median and either the minimum or maximum value for variables that were not normally distributed. In the case of qualitative variables, we calculated frequencies and proportions.
The frequencies among the risk groups were determined using X2, Fisher’s exact, and Kruskal–Wallis tests according to the type of variable and its distribution. We presumed that p < 0.05 indicated statistical significance. We also used a Cox proportional hazard model with mortality as the event. Overall and event-free survival was analyzed using the Kaplan–Meier method, and a log-rank test was used to assess whether there were significant differences between the groups.
Results
Clinical features
From January 2010 through December 2018, 253 patients with LCH were diagnosed at the 19 participating institutions (Table 2). The median age at diagnosis was 19 months. Patients in Group 1 had a mean age at diagnosis of 15 months (range 0.75-120 months), patients in Group 2 had a mean age at diagnosis of 25.5 months (range 0.62-192 months), and patients in Group 3 had a mean age at diagnosis of 28 months (range 0.2-180 months) (p = 0.0001). The male-female ratio of the total cohort was 1.3:1. The most frequently affected sites were the bone (178 cases; 70%) and the skin (131; 51.7%). Table 3 lists the primary clinical manifestations at diagnosis.
Table 2 Age and reference hospital
| Variable | Group 1 (n = 112) | Group 2 (n = 71) | Group 3 (n = 70) |
|---|---|---|---|
| Age (years) | Median (min-max) 1.2 (0.06-10) | Median (min-max) 2.12 (0.05-16) | Median (min-max) 2.3 (0.01-15) |
| Reference hospital | n (%) | n (%) | n (%) |
| Instituto Nacional de Pediatría | 23 (20.5) | 21 (29.5) | 6 (8.57) |
| Hospital Infantil Teletón de oncología | 1 (0.89) | 4 (5.6) | 2 (2.85) |
| Centro Médico Nacional 20 de noviembre | 2 (1.78) | 6 (8.4) | 4 (5.71) |
| Hospital Ángeles del Pedregal | 1 (0.89) | 1 (1.40) | 0 (0) |
| Centro Estatal de Oncología de Campeche | 2 (1.78) | 3 (4.22) | 1 (1.42) |
| CECAN, Veracruz | 7 (6.25) | 2 (2.81) | 3 (4.28) |
| Hospital General de Celaya | 1 (0.89) | 2 (2.81) | 1 (1.42) |
| Hospital para el niño del IMIEM | 24 (21.4) | 2 (2.81) | 4 (5.71) |
| Hospital Materno Infantil del ISSEMyM | 2 (1.78) | 2 (2.81) | 3 (4.28) |
| Hospital General de México | 2 (1.78) | 2 (2.81) | 0 (0) |
| Hospital General de León | 5 (4.46) | 1 (1.40) | 2 (2.85) |
| Hospital General "Dr. Agustín O’Horán" | 8 (7.14) | 9 (12.6) | 7 (10) |
| Hospital para el Niño Poblano | 2 (1.78) | 2 (2.81) | 0 (0) |
| Hospital del Niño "Federico Gómez Santos," Saltillo | 2 (1.78) | 0 (0) | 0 (0) |
| HNRNP Tabasco | 3 (2.67) | 0 (0) | 2 (2.85) |
| Hospital Infantil de Morelia | 2 (1.78) | 1 (1.40) | 3 (4.28) |
| Hospital Infantil de México | 22 (19.6) | 9 (12.6) | 26 (37.14) |
| Hospital Juárez de México | 0 (0) | 1 (1.40) | 3 (4.28) |
| Hospital Universitario UANL | 3 (2.67) | 3 (4.22) | 3 (4.28) |
Table 3 Clinical characteristics by group in patients with LCH according to the risk group at diagnosis
| Variable | Group 1 Frequency (%) (n = 112) | Group 2 Frequency (%) (n = 71) | Group 3 Frequency (%) (n = 70) | p-value |
|---|---|---|---|---|
| Gender | ||||
| Male | 58 (0.51) | 47 (0.66) | 40 (0.57) | 0.15 |
| Female | 54 (0.48) | 24 (0.33) | 30 (0.42) | |
| Clinical manifestations | ||||
| Skin lesions | 69 (0.61) | 42 (0.59) | 19 (0.27) | 0.000 |
| Adenopathy | 76 (0.67) | 33 (0.46) | 18 (0.25) | 0.000 |
| Diabetes Insipidus | 14 (0.12) | 10 (0.14) | 12 (0.17) | 0.68 |
| Lung involvement | 16 (0.14) | 5 (0.07) | 0 (-) | 0.003 |
| Bone lesions | 66 (0.58) | 56 (0.8) | 55 (0.78) | 0.002 |
| Otitis | 15 (0.13) | 10 (0.14) | 1 (0.01) | 0.016 |
| Weight loss | 24 (0.21) | 6 (0.08) | 3 (0.04) | 0.002 |
| Proptosis | 7 (0.06) | 12 (0.17) | 7 (0.10) | 0.07 |
| General symptoms | 65 (0.58) | 20 (0.28) | 15 (0.21) | 0.000 |
| Tumor | 1 (0.009) | 2 (0.02) | 0 (-) | 0.28 |
| Retinal gliosis | 0 (-) | 1 (0.01) | 0 (-) | 0.27 |
| Medullary Syndrome | 0 (-) | 0 (-) | 1 (0.01) | 0.26 |
| Pain | 0 (-) | 2 (0.02) | 1 (0.01) | 0.22 |
| Vasculitis | 1 (0.008) | 0 (-) | 0 (-) | 0.53 |
| Diarrhea | 0 (-) | 1 (0.01) | 0 (-) | 0.27 |
| Gingival hyperplasia | 1 (0.008) | 0 (-) | 0 (-) | 0.53 |
| Hematopoietic condition | 89 (0.79) | 0 (-) | 0 (-) | 0.000 |
Test statistic = X2.
The most frequent sites of bone involvement were the skull (127; 50.1%), spine (32; 12.6%), femur (28; 11%), pelvis (14; 5.5%), and humerus (11; 4.3%).
The most frequently positive immunohistochemistry markers for diagnosis included CD1a (199; 78.6%), S100 (166; 65.6%), CD207 (71; 28%), CD68 (41; 16.2%), and Birbeck granules, determined by electron microscopy (16; 6.3%).
Pulmonary involvement was observed in 13 patients (11.6%) in Group 1 and 5 patients (7%) in Group 2, mainly presenting as alveolar infiltrate in 16 cases (8.7%), bullae in one case (0.54%), and bronchiectasis in one case (0.54%).
Forty-eight patients (42%) in Group 1 had bone marrow involvement, 62 (53%) had splenomegaly, and 39 (34.8%) had liver involvement. The values of hemoglobin, leukocytes, and platelets at diagnosis were statistically significantly different among the patient groups. Platelet levels (median 204,000 cells/μL) and hemoglobin (median 7.6 g/dL) were lower in patients in Group 1 (p = 0.0001).
Treatment
Two hundred and thirty-five patients (92.5%) received treatment according to the national protocol based on the LCH III protocol with 6 mg/m² vinblastine and 40 mg/m² prednisone for 6 or 12 months, depending on the risk group. Ten patients (3.9%) received first-line treatment with vincristine (1.5 mg/m²), cytarabine (100 mg/m²), and prednisone (20 mg/m²). Six patients (2.3%) received the LCH II protocol, three patients (1.1%) received treatment with etoposide 100 mg/m² as a single drug, and five patients (1.9%) underwent surgical treatment only. Three patients (1.1%) received no treatment.
Of the patients who underwent chemotherapy, 61.2% had a complete response with the first induction; this rate increased to 87% after the second induction treatment. Nineteen patients (7.9%) had no response or experienced disease progression after the second induction; they switched to a second-line of treatment. That treatment included cytarabine (100 mg/m²), prednisone (40 mg/m²) or cladribine (5 mg/m²), and cytarabine 100 mg/m² (Table 4).
Table 4 Treatment response, relapse and deaths by groups
| Variable | Group 1 Frequency (%) (n = 112) | Group 2 Frequency (%) (n = 71) | Group 3 Frequency (%) (n = 70) | p-value |
|---|---|---|---|---|
| 1st induction | 109 | 65 | 66 | |
| Complete response | 40 (0.35) | 41 (0.57) | 53 (0.75) | 0.000 |
| Very Good partial | 36 (0.32) | 14 (0.19) | 10 (0.14) | 0.016 |
| Partial | 11 (0.09) | 12 (0.16) | 6 (0.08) | 0.25 |
| No response | 12 (0.10) | 4 (0.05) | 1 (0.01) | 0.03 |
| Progression | 12 (0.10) | 0 (-) | 0 (-) | 0.00 |
| 2nd induction | ||||
| Complete response | 84 (0.75) | 61 (0.86) | 64 (0.91) | 0.012 |
| Very good partial | 13 (0.11) | 4 (0.05) | 2 (0.028) | 0.092 |
| Partial | 2 (0.017) | 1 (0.014) | 0 (-) | 0.791 |
| No response | 6 (0.05) | 1 (0.014) | 3 (0.02) | 0.391 |
| Progression | 4 (0.035) | 3 (0.042) | 2 (0.028) | 1.0 |
| Relapse | 23 (20.5) | 10 (14.1) | 3 (4.28) | 0.009 |
| Death | 19 (16.9) | 2 (2.81) | 1 (1.42) | 0.000 |
Test statistics: Fisher×s exact test.
Relapse
Thirty-six patients (14.2%) experienced disease relapse after presenting complete remission. The most frequent sites of relapse included the skin, bone, lymph nodes, and liver (Table 5). The second-line treatments consisted of several combinations: (1) cytarabine, vincristine, and steroids in 14 cases (38.8%); (2) vinblastine and prednisone in 11 cases (30.5%); (3) clofarabine in one case (2.7%); (4) cladribine in four cases (11.1%); (5) etoposide and cytarabine in five cases (13.8%); and (6) no treatment in one case (2.7%).
Table 5 Relapse sites in patients with LCH according to the group
| Variable | Group 1 Frequency (%) (n = 112) | Group 2 Frequency (%) (n = 71) | Group 3 Frequency (%) (n = 70) | p-value |
|---|---|---|---|---|
| Skin | 3 (0.02) | 2 (0.02) | 3 (0.04) | 0.900 |
| Bone | 10 (0.08) | 8 (0.11) | 0 (-) | 0.007 |
| Lymph nodes | 5 (0.04) | 0 (-) | 1 (0.01) | 0.14 |
| Liver | 3 (0.02) | 1 (0.01) | 4 (0.05) | 0.38 |
| Spleen | 2 (0.01) | 1 (0.01) | 2 (0.02) | 0.73 |
| Lung | 1 (0.008) | 0 (-) | 1 (0.01) | 0.75 |
| Middle ear | 1 (0.008) | 0 (-) | 0 (-) | 1.0 |
| Orbit | 1 (0.008) | 0 (-) | 1 (0.01) | 0.75 |
| Pituitary gland | 0 (-) | 1 (0.01) | 0 (-) | 1.0 |
| Bone marrow | 0 (-) | 1 (0.01) | 0 (-) | 1.0 |
Test statistic: Fisher×s exact test; LCH: langerhans cell histiocytosis.
Twenty-two patients (8.6%) died. The causes of death included refractory disease in six patients (27.2%), septic shock in 10 patients (45.4%), infection in two patients (9%), hematological alteration in one patient (4.5%), pneumonia in one patient (4.5%), and intracranial hemorrhage in one patient (4.5%).
The overall survival rate was 91.3%, and it was lower for patients in Group 1 (77%) compared with patients in Groups 2 (97%) and 3 (100%) (p = 0.001; Fig. 1). Event-free survival was 66% for patients in Group 1, 74% for patients in Group 2, and 93% for patients in Group 3 (p = 0.0012; Fig. 2).
The mortality risk of each independent variable was analyzed by calculating the hazard ratios (HR) with a 95% confidence interval (CI); statistically significant differences were incorporated into a broad multivariate Cox proportional hazard model. We then used the stepwise backward method to reduce those differences until the best-fit model was obtained (log-likelihood = −25.94, X2 (3) = 33.83, p = 0.000). Finally, the reduced model was verified using tests and diagnostics for a proportional hazard model with p = 0.10. In this model, variables associated with mortality included an age under 1 year (HR 5.037, p = 0.001), splenomegaly (HR 6.494, p = 0.000), and refractory disease (HR = 2.80, p = 0.005) tables 6 and 7.
Table 6 Cox proportional hazards model for mortality in pediatric patients with LCH in Mexico (2010-2018) n = 253
| Variable | HR (IC95%) | p-value |
|---|---|---|
| Sex (female) | 1.041 (0.43-2.47) | 0.92 |
| Risk Group | ||
| Risk group 1 | 11.985 (2.78-51.66) | 0.001 |
| Risk group 2 | 0.292 (0.06-1.25) | 0.09 |
| Relapse | 2.119 (0.85-5.25) | 0.10 |
| Clinical manifestations | ||
| General symptoms | 1.709 (0.72-4.02) | 0.22 |
| Dermatitis | 4.160 (1.39-12.36) | 0.010 |
| Lymphadenopathy | 2.088 (0.84-5.17) | 0.11 |
| Diabetes insipidus | 1.382 (0.46-4.10) | 0.56 |
| Lung disorders | 2.584 (0.86-7.67) | 0.08 |
| Otitis | 1.586 (0.46-5.39) | 0.46 |
| Proptosis | 1.066 (0.24-4.58) | 0.93 |
| Splenomegaly | 6.821 (2.75-16.9) | 0.000 |
| Bone marrow infiltration | 4.669 (1.98-10.99) | 0.000 |
| Bone lesions | 0.246 (0.10-59.4) | 0.002 |
| Laboratory studies | ||
| Hyperbilirubinemia | 6.780 (2.73-16.83) | 0.000 |
| Hypoproteinemia | 6.712 (2.59-17.37) | 0.000 |
| Hypoalbuminemia | 4.466 (1.80-11.07) | 0.001 |
| Hemoglobin (mg/dL) | 0.734 (0.62-0.86) | 0.000 |
| Leukocytes | 0.983 (0.91-1.06) | 0.66 |
| Platelets | 0.993 (0.990-0.996) | 0.000 |
HR: hazard ratios.
Table 7 Reduced COX model
| Variable | HR (IC95%) | p-value |
|---|---|---|
| Age < 1 year | 5.037 (1.90 a 13.31) | 0.001* |
| Splenomegaly | 6.494 (2.48 a 16.96) | 0.000* |
| Refractory disease | 2.802 (1.00 a 7.84) | 0.050* |
Log likelihood: −90.17; X2 (3): 33.87; p = 0.000*; HR: hazard ratios.
The primary factors associated with mortality included belonging to Group 1 (HR 8.23; 95% CI 2.4-27.9; p = 0.001) and the involvement of the skin, bone, liver, spleen, or bone marrow.
Discussion
LCH is a rare disease; its incidence in Mexico is similar to that reported in other countries. The aberrant differentiation of mononuclear cells through the mitogen-activated protein kinase pathway characterizes LCH and conditions its activation23.
We have reported the results of the largest Mexican cohort to date (253 patients) diagnosed with LCH over 8 years at 19 institutions. Roughly half of the patients (44%) had high-risk organ disease. Patients in this group exhibited a survival rate of 77%, similar to that reported internationally (1 year of overall survival up to 99%)7-24.
Validated risk-stratification criteria for children include disease sites and responses to initial therapy. Patients with lesions in "risk organs," including the bone marrow, spleen, or liver (Group 1), have a significantly higher risk of mortality than patients with lesions limited to "non-risk" sites (7). Risk stratification for LCH is based on the analysis of outcomes of prospective pediatric trials. Patients with high-risk LCH exhibit a survival rate of almost 90%. Still, outcomes are significantly worse if the disease progresses during the 1st 12 weeks of therapy8.
In patients without risk organ involvement, the survival rate has been reported to be 98.2%, with a relapse rate of 26% for patients in Group 225. We noted similar results in our population.
Gene alterations, such as the BRAFV600E mutation, have been shown in recent years to be a factor associated with high-risk characteristics, in addition to relapses and disease refractoriness26. However, it is not possible to routinely test for gene alterations in our population; this fact is a limitation of our study.
In general, the treatment received by our patients is recommended by the Histiocyte Society based on vinblastine and prednisone, which has been the standard treatment used internationally27. The treatment duration is 12 months for patients with risk organ disease and 6 months for those without organ disease. The primary second-line treatments used in patients with relapsed or refractory disease include cytarabine, etoposide, steroids, and other drugs28,29. The LCH-IV study is one of the largest LCH trials to date; the Histiocyte Society is currently carrying it out. This prospective multi-arm study will evaluate first- and second-line treatments and include patient follow-up30. There is no question that multisystem/multifocal bone involvement constitutes a recurring problem31. Our patients could access drugs, such as cladribine, which is currently recognized as yielding adequate outcomes.
Single bone lesions are effectively treated with limited curettage or corticosteroid injections32. Large pelvic or vertebral lesions not amenable to curettage may be treated with radiation therapy. Our patients were offered a similar chemotherapy regimen as children in Group 1 who relapsed. Recently, indomethacin has been considered an alternative for bone LCH33.
Besides our inability to detect BRAF V600E, another limitation of this retrospective study was that the histopathological study and follow-ups were performed according to the protocols of each center.
Conclusion
This study represents the first effort in Mexico to conduct a national analysis of LCH. We observed that LCH is a pathology characterized by relatively homogeneous patient care across Mexico. It has been reported that survival is acceptable; however, the frequency of recurrence is still high. Therefore, increased access to second-line drugs with limited toxicity is necessary. The challenge now is to access BRAF V600E testing so that patients can benefit from effective targeted therapies and form a cooperative group that involves most institutions.
