2.1 Study procedure and data acquisition

This was an observational retrospective cohort study. All patients in Sweden diagnosed with CLL according to the World Health Organization criteria and treated with idelalisib after its approval in March 2014 were included. A preliminary survey revealed that from the end of 2017 when venetoclax became available and onwards the use of idelalisib became minimal. The inclusion period was therefore limited to 2.5 years, that is, from approval in March 2014 until September 2017, which also allowed long-term follow-up of our cohort. Ethics committee approval was obtained before commencement of the study (Dnr: 2015/1025-32). As this was a retrospective observational study no informed patient consent was required. The study was performed in accordance with the ethical principles of Declaration of Helsinki and in compliance with national laws.

Patients treated with idelalisib were identified from the Swedish pharmacy registry and cross-checked with the patient file via the regional representatives of the Swedish CLL study group. Medical files were reviewed in detail from the date of diagnosis until patient's death or until the end of data collection defined as the last follow-up. Data were abstracted to case record forms (CRFs).

The Cumulative Illness Rating Scale (CIRS)¹³ was applied to assess comorbidity other than CLL at the start of treatment. Response to treatment as well as hematological toxicity was evaluated according to iwCLL guidelines.¹⁴ Major infections and hospitalization that occurred during treatment and until 3 months after treatment termination were recorded as well as use of prophylactic antibiotics. Detailed information about treatment-related adverse events (colitis, transaminitis, and pneumonitis) was recorded.

Data from CRFs were transferred to a database and the information was systematically cross-checked and validated for accuracy, along with additional validation during the statistical analysis.

2.2 Statistical methods

Descriptive statistics such as counts, percentages, medians, and range were used to summarize categorical and continuous data. Distributional differences between groups were tested using Fisher's exact test for categorical data and the Mann–Whitney or the Kruskal–Wallis test for continuous data. The proportion of responses is presented together with exact (Clopper–Pearson) 95% binomial confidence intervals. Time in the analysis of PFS was calculated from date of treatment start to date of progression, date of treatment change or date of death, whichever came first. Length of overall survival (OS) was calculated from the date of treatment start to the date of death. For event-free patients or patients still alive, time was calculated from treatment start to the last clinical visit date.

Survival curves were calculated using the Kaplan–Meier method and differences between groups in time to failure/death were tested using the log-rank test. All statistical analyses were performed using the software Stata (version 17).

3.1 Patients

Thirty-seven patients were identified at 17 Swedish centers until September 2017. Baseline characteristics are shown in Table 1. The median age was 69 years (range 50–89). Most patients had a good performance status (54% had Eastern Cooperative Oncology Group [ECOG] 0–1) and few comorbidities (78% had CIRS score ≤6). Seventy-three percent had Rai stage III/IV. Nineteen patients (51%) had del(17p) or TP53 mutation. Additionally, five patients (14%) had del(11q). The median number of prior therapies was 3 (range 1–11). Eleven patients (30%) had previously received ibrutinib treatment.

3.2 Length of therapy and efficacy

All patients received idelalisib at a starting dose of 150 mg twice daily with (n = 28, 76%) or without (n = 9, 24%) concomitant rituximab as decided by the responsible physician. Idelalisib was continued until progression or non-tolerability. The median length of therapy was 9.8 months (range 0.9–44.8). The median time between stop of idelalisib and start of next treatment was 16.5 months (range 4.6–46.9).

The ORR was 65% (24/37): 21 partial responses (PRs) and 2 PR with lymphocytosis (PR-L) and 1 complete remission (CR). The response rates were similar in men and women and neither age nor del(17p)/TP53 mutation appeared to affect response rates. Median time to response was 5.7 months (range 1.5–22.9). Response rate was similar irrespective if patients had received prior ibrutinib (n = 11) or were BTKi naïve (n = 26) (ORR 64% and 65%, respectively). Seven out of 9 patients (78%) responded to single-agent idelalisib and 17 of 28 patients (61%) to idelalisib in combination with rituximab.

Median follow-up was 33.6 months (range 10.8–57.6) and the median PFS was 16.4 months (95% CI: 10.4–26.3) (Figure 1A). None of the baseline variables (gender, age, clinical stage, CIRS, del(17p)/TP53, del(11q), or number of previous treatment lines [1–2 and more than 3]) were significantly associated with PFS. Median OS had not been reached; the survival rate at 24 months of follow-up was 75% (95% CI: 57%–86%) (Figure 1B). OS was significantly associated with number of previous treatment lines (1–2 and more than 3) (p = .01) (Figure 2) but not with sex, age, stage, CIRS, or cytogenetics.

At 24 months of follow-up, 27% of ibrutinib pre-treated patients remained progression-free (95% CI: 7%–54%) compared to 45% of patients who were BTKi naïve (95% CI: 25%–63%) (not significant).

Also measured at 24 months of follow-up 53% receiving single-agent idelalisib remained progression-free (95% CI: 18%–80%) compared to 36% of patients who received the drug in combination with rituximab (95% CI: 19%–53%) (not significant).

3.3 Toxicity

Toxicity is listed in Table 2. Side effects (≥grade 3) occurred in 10/37 patients and included diarrhea/colitis (n = 7), transaminitis (n = 1), rash (n = 1), and pneumonitis (n = 1). Reasons for discontinuation of idelalisib were colitis (n = 8), progressive CLL (n = 7), grade 3–4 infection (n = 4), Richter transformation (n = 3), pneumonitis (n = 2), allogeneic stem cell transplantation (n = 1), mucositis (n = 1), and liver failure (n = 1). Eighty-nine percent of patients (33/37) were hospitalized during idelalisib therapy. Of those hospitalized, infection (n = 24) was the most common reason followed by AE (n = 12). Other reasons for hospitalization (different reasons for all patients) occurred in 10 patients. The most common ≥grade 3 infections were pneumonia (n = 9), febrile neutropenia (n = 4), septicemia (n = 3), and opportunistic infections (n = 5, including systemic fungal infection [n = 1] and pneumocystis pneumonia [n = 4]). Sixty percent of patients received concomitant antibiotics (all were Pneumocystis jirovecii pneumonia prophylaxis). Neither age nor CIRS, ECOG or number of previous treatment lines (1–2 and more than 3) were significantly associated with the risk of infections. The most common AEs were colitis (n = 6), diarrhea (n = 3), GI-bleeding (n = 1), heart failure (n = 1), and mucositis (n = 1).

Neutropenia ≥grade 3 occurred in 24 of the patients and thrombocytopenia in 15 patients but no anemia ≥grade 3 was observed. Other malignancy occurred in one patient (squamous cell carcinoma in the skin with brain metastasis).

Salvage therapy after idelalisib treatment was started in 26 patients. Regimens used were: ibrutinib (n = 15), venetoclax (n = 4), CHOP-based therapies (n = 4, all but one had Richter transformation), bendamustine + ibrutinib (n = 1), and two patients received allogeneic stem cell transplantation.

Seventeen patients have died and causes of death were disease progression (n = 12), concomitant colitis and septic shock (n = 1), hemophagocytic syndrome (n = 1), myocardial infarction (n = 1), other malignancy (n = 1), and liver cirrhosis (n = 1).