RO + pediatric LCH, carries a bad prognosis in patients failing the 1st line treatment [3]. Patients were divided into two groups according to the period and the treatment they received. Those through LCH III protocol, who received treatment before mid-2012 and having ID MTX during induction but without salvage by 2-Cda. Otherwise, those through LCH IV protocol who received treatment after 2012 without ID MTX in induction but with 2-CdABR salvage treatment; a regimen that has been proved to be effective, but toxic [22]. Nowadays as RO + LCH frequently harbor the BRAFV600E mutation [24], the innovative anti B RAF targeted therapy is becoming in common use in refractory RO + [13, 25]. However, being an off-label, expensive drug with a higher reactivation risk at discontinuation, anti B Raf is becoming a useful option as a 3ed line treatment [25].

ID MTX in LCH III protocol did not prove any effectiveness on survival on the long term, reason for which it has been omitted afterwards in the subsequent LCH IV protocol [4]. In our study, there was a rise in the cumulative incidence of DP in patients not receiving ID MTX with a tendency to be statistically significant. The present updated analysis with 67 patients is more prominent than our previous one which comprised a smaller number of 50 patients [21]. Since pediatric LCH is a rare disease, it is difficult to statistically empower such a comparison at a reasonable alpha error rate. On the other hand, ID MTX was associated with better EFS with statistically significant results, confirming our previous study and others [21, 26].

Failing induction DP RO + comprise most early deaths [6, 22, 27]. Otherwise, disease REA, although rare in RO + , is still responsible of mortality in a small number of patients [28]. In our study, DP was accompanied by the worst OS in relation to REA p < 0.001. This is even worse if REA occurred in RO + organs p 0.008, reflecting the benignity of RO- responding to repetition of 1st line treatment. While 2-CdABR was provided starting from 2012, OS in DP with 2-CdABR was 56% vs 8% without p 0.004. This statistically proves the high regimen efficiency with the best response in patients with hematopoietic system involvement. All these patients showed better status with 2-CdABR at last follow up, differently from hepato-splenic DP that responded partially to 2-CdABR. This could be partially explained by the presence of a resistant sclerosing cholangitis. Donadieu et al. showed the prompt effect of 2 CdA aracytine in a refractory RO + cohort -excluding sclerosing cholangitis- in a phase II LCH-S-2005 study [22]. It showed an overall response rate (ORR: NAD and ADB) of 92% and a 5-year OS 85% contrarily to a previous study LCH-S-98 study revealing the lack of disease control by 2-CdA monotherapy with a 2-year OS of 48.0% [10]. In an updated nationwide survey from 2005 to 2019 in Japan, Tanigushi et al. showed 21 RO + patients failing 1st line, where the ORR was 50% the OS of 86% and EFS of 77% [12].

In our series, the role of 2-CdABR could not be proven in REA with even a worse survival than other salvage lines; this probably due to considerable REA on RO- mode responding to repeating 1st line treatment. In REA on RO + mode, neither 2-CdABR or other chemotherapy salvage regimen could control the disease promptly. In this situation 2-CdABR is even too toxic in this hepatic sclerosing cholangitis form; a more severe resistant disease presenting at a later stage reaching biliary cirrhosis and hepatic cell failure beyond the scope of chemotherapy. Otherwise, cytopenia and splenomegaly as a RO + mode of reactivation, responded exceptionally to both 2-CdABR and other salvages; but these results remain subjective. Outcome of 2- CdA on REA remains variable as Imamura showed a limited outcome if REA occurred on RO + mode [11] contrarily to the possibility of rescuing RO + twice, the first on REA RO + liver and spleen and the other on REA RO- [29].

In our series, 2-CdA based regimen was responsible of severe infectious complications mainly in those with DP, showing pneumonia in 1/3 of them. Moreover, 2 patients out of 3 died from infection in a better status as regard the initial disease. Otherwise infection was exceptional in REA -in the form of proctitis and pancreatitis- and was of favorable outcome. As 2-CdA causes lymphopenia, and prolonged pancytopenia [30], Donadieu et al. showed that 2- CdA induced immune suppression caused half of the deaths from fulminant viral infection [22].

The incidence of cancer deaths is 70% in low/middle-income countries (LMICs) [31]. This is partly due to the lack of chemotherapeutic agents [14]. In this optic, we investigated the outcome of RO + for proper treatment planning according to cost effectiveness. In our early practice, between 2007 and 2012 we could not use 2- CdA, an expensive drug lacking the evidence, and this led to the mortality of all RO + failing 1st line LCHIII protocol. Between 2012 and end 2019, 2-CdABR rescued all DP RO + patients with hematopoietic involvement and most of hepatosplenic ones. JSLG 02 protocol offers a salvage treatment including cyclophosphamide, doxorubicin, Vincristine, prednisone cyclosporin A in RO + patients failing 1st line treatment. However, its results are non-satisfactory with survival reaching 70% requiring the use of further 2-CdA or targeted therapy [32]. In low-income countries, further salvage treatment beyond first-line therapy for LCH remains economically difficult. Therefore, to improve survival, reducing 1st line treatment failure remains strategic by augmenting induction and prolonging maintenance in RO + patients. Narula et al., showed that RO + LCH receiving oral etoposide augmented induction and maintenance had early and durable responses. Prolonging maintenance including methotrexate lowered reactivation rates in RO + and RO − LCH, resulting in excellent survival and reduced the need for salvage [33]. In our population, it is still hard to conclude the superiority of ID MTX with confidence due to the unbalanced 2 subgroups of a relatively small population. This would invite the re discussion of the role of ID MTX in decreasing DP at end of induction. If ascertained, this would reduce the need for 2- CdA as a salvage line and thus offer favorable cost-effective management for a potentially lethal pediatric RO + LCH.