Patients
In total, 20 patients were enrolled (median age: 43 years, range: 19–77 years) between November 2013 and November 2015 from eight hospitals of the Consortium for Improving Survival of Lymphoma (CISL). However, the central pathology review during the final analysis reported that one reported case of HL was actually EBV-associated lymphoproliferative disease. Thus, we analyzed 13 patients with HL and six patients with PMBCL. All patients with HL initially received ABVD chemotherapy whereas all but one of the patients with PMBCL received R-CHOP chemotherapy as first-line treatment. The demographic and clinical characteristics of the patients with HL and PMBCL at the time of enrollment did not differ except for the number of patients with stage IV disease; 84% (16/19) of the patients had refractory disease (Figure 1A).
Response to ruxolitinib
Among 13 patients with HL, five patients achieved PR and one patient showed CR (Figure 1B). The other patient with HL achieved SD after the second cycle of treatment. He showed clinically improvement of symptoms and disappearance of tumor lesions although some lesions were persistent (Figure 1C). The median duration of response was 5.6 months in six responders including one CR and five cases with PR (95% confidence interval [CI]: 0.01–12.1 months). However, all PMBCL patients rapidly progressed after the first or second cycle of treatment (Figure 2). The 74-year old female patient achieving CR had stage III, lymphocyte-rich HL at the time of enrollment (Patient number 18, Figure 2). Although she had experienced two episodes of relapse after she was initially treated with ABVD, she maintained her CR status until the 15th cycle. Of the three patients with HL who had previously received brentuximab vedotin, one patient achieved a PR and maintained this response until the tenth cycle (Figure 2). As a result, the overall disease control rate for HL was 54% (7/13) although the overall disease control rate for all participants was 36.8% (7/19).
JAK2 amplification
The waterfall plot of this study demonstrated all of the responders to ruxolitinib were HL patients (Figure 3A). FISH analysis of JAK2 amplification was performed in nine patients for whom tissue samples were available to analyze the association of JAK2 amplification with response to ruxolitinib. Out of nine patients, JAK2 amplification was demonstrated in six patients (four HL, two PMBCL, Figure 3B, C). Three of these six patients, all of whom had HL, showed PR or SD. However, the three HL patients without JAK2 amplification showed PD (Figure 3D). Although the number of patients was too small for statistical analysis, HL patients with JAK2 amplification seemed to have a high probability of responding to ruxolitinib, while the two PMBCL patients with JAK2 amplification did not respond to ruxolitinib (Figure 3D).
Survival outcome and toxicities
The median PFS of HL patients (3.6 months, 95% CI: 1.4–5.8 months) was longer than that of PMBCL patients (0.9 months, 95% CI: 0.72–1.08 months, Figure 4A). The median OS of HL patients was not reached within the median follow-up of 37.0 months (95% CI: 32.5–41.5 months), and was thus also much better than that of PMBCL (3.0 months, 95% CI: 0.0–12.6, Figure 4B). Treatment-related adverse events were reported in 14 patients (73.6%), however, most events were grade 1 or 2 (Figure 4C). Grade 3 neutropenia and anemia were observed in three patients, all of whom recovered. As a result, there was no dose-modification according to the occurrence of hematologic and non-hematologic toxicities.