To our knowledge, this is the first study discussing a novel LCH grouping system and treatment protocol based on the severity of patients’ clinical manifestation of inflammatory or malignant symptoms. The overall 4-year OS and EFS have been the highest compared to the other published data in LCH cohorts, albeit with a relatively small patient number in this pilot study.

Guangzhou Women and Children’s Medical Center has started pediatric LCH diagnoses and treatment since the 1990s, with anually averaged number of 30 to 40 LCH cases. For years, we have observed an interesting phenomenon during the clinical practice of LCH treatment under LCH III or JLSG-96. When LCH is mainly manifested as inflammatory symptoms, such as systemic rash, multiple bone lesions, pus and exudate from the outer ear, long-term diarrhea, pneumonia, etc., the disease progresses sluggishly even with MS involvement, and the severity is normally of low-grade. Once malignant symptoms appear, such as long-term fever, splenomegaly, anemia, thrombocytopenia, and/or mediastinal mass, the disease would progress rapidly, severely or even fatally although fewer systems involved and the rash was normally mild.

During the 30 years of LCH treatment practices, we have also observed that LCH patients are with varied response to different drugs, suggesting the currently world-known grouping system and treatment protocol based only on the number or risk of organs involved might be insufficient. Based on the observed phenomena, we grouped patients according to their symptoms while scoring systems and disease status assessments have also been adjusted. In the other grouping systems, at least bi-lineages of the blood system should be affected when judged as blood system involvement, whereas in this study, patients showed continued anemia before or after treatment (excluding chemotherapy-induced or malnutritional reasons) even without leukopenia or thrombocytopenia were still judged as blood system involvement. When conducting liver evaluation, the method of judging the liver size (liver subcostal distance) was discarded, and the Child–Pugh liver function evaluation method was adopted. In terms of bone damage evaluation, not only the situation of special sites such as anterior bones, skull base bones, and vertebrae were considered, but also whether the affected bones are functional bones or load-bearing bones were taken into account in this protocol. The involvement of these bones could affect the patients’ quality of life, which is sometimes overlooked but highly recommended to be observed in previous classical LCH protocols. To reduce neurotoxiciy and cardiotoxicity, we used new alkaloid of VDS and THP, and reduced the dose of THP to 20 mg/m2 to mitigating bone marrow suppression.

After grouping stratification according to the presentation of inflammatory or malignant symptoms, patients in the inflammatory symptom group who were treated with VDS + PSL only, have achieved good long-term EFS without adding vemurafenib, regardless of the presence of BRAF mutation. Only one patient with CEP72 T/T genotype developed drug-related intestinal obstruction, which resolved within a short time after discontinuation of VDS. The patient was treated with reduced doses of VDS in the remaining phase of regimen without reoccurrence of intestinal obstruction. No significant adverse effects occurred in any of the remaining cases in Group A. This protocol demonstrates a novel treatment strategy with low chemotherapy toxicity, good tolerability, short hospitalization days and low cost.

Patients in the malignant symptom group, given intensive treatment with multiple drugs at the beginning, were able to control the disease progression rapidly, and all of them could reach NAD after six weeks of initial treatment, and then entered the maintenance phase of treatment, which also could achieve good long-term EFS. This protocol reduced the dose of anthracyclines, with a cumulative THP dose of only 160 mg/m2 (compared to the cumulative ADR of 210 mg/m2 in JLSG-96). Cardiac function and ECG were routinely monitored during the course of treatment, and no abnormalities were found. Most of the treatment-related adverse effects were alopecia and hematological toxicity. As the majority of patients in Group B have reached grade 4 (CTCAE 5.0) hematologic toxicity in the third week of this regimen, we effectively reduced the incidence of infectious complications by temporarily suspending the treatment. After the neutrophil count returned to 0.5 × 109/L or above, the regimen was continued.

The time of first-episode diabetes insipidus in Group A was significantly earlier than in Group B, approximately 1 year earlier. This may be related to the involvement of the skull base bones in Group A patient, which needs to confirm by more cases, however, it reminds us that we should pay more attention to the urine volume and water intake of patients with skull base bones involvement during the follow-up. The event onset time of the two cases in Group B was after 18 months since treatment (6 months after drug withdrawal), which suggests a 1-year maintenance treatment might not be sufficient and longer duration of maintenance treatment should be considered in the future.

During the course of treatment, in both Groups A and B, we found that liver symptoms such as hepatomegaly, elevated liver enzymes and elevated bilirubin would be present for a long period in some patients. Nonetheless, the liver function assessment results remained at Grade A or B according to Child–Pugh classification, which did not lead to disease progression or affect their long-term survival. Additionally, we found glucocorticoid combined with low cytotoxic agents (e.g., VDS) could significantly slow down the progression of biliary sclerosis without intensive chemotherapy for hepatic symptoms only during LCH treatment. Our findings suggested that in pediatric LCH patients only presenting inflammatory symptoms, regardless of manifestation of MS or RO+ involvement, would still achieve reasonable disease control and/or good clinical outcomes with less intensive treatment of VDS + PSL.