This study was approved by Ethics committee of Wroclaw Medical University. Forty-two CALR-positive ET (n = 28) and MF (n = 14) patients underwent targeted sequencing. The median age at diagnosis of the entire patient series was 58 years (range, 24–87); 20 were males (47.6%). Due to the fact that the research material was archival and came from the years 2015–2020, in Table 1, we present clinical and biological information about patients for whom we were able to obtain a medical history. Each group was further divided according to the type of CALR mutation. In the study group, we had patients only with a 52-bp deletion (type-1 mutation) and a 5-bp insertion (type-2 mutation) (Appendix 1). Type 1 CALR mutation occurred in 54% (n = 15), and type 2 occurred in 46% (n = 13) of patients with ET. In MF, type 1 definitely predominates and occurred in 93% (n = 13) of cases, only one patient had a type 2 mutation. Additional mutations were detected by the NGS sequencing panel in 48% of patients in the whole cohort (Appendix 2). The analysis detected 44 pathogenic, likely pathogenic, or uncertain significance variants. According to scientific reports, we reported variants on VAF above 3%. The most frequently mutated genes in the study population were ASXL1, TET2, and DNMT3A, which are largely associated with epigenetic regulatory mechanisms (Fig. 2). Twelve variants were detected in the ASXL1 gene, including one patient with 2 co-occurrent variants, and one patient with three co-occurrent variants. Fourteen patients had two or more than two additional mutations (33%). Mutations in genes: ASXL1, SRSF2, EHZ2, IDH1, IDH2, U2AF1, defined as high molecular risk mutations (HRM) had ten patients. Seven patients transformed from ET to MF. Among these patients, five had additional mutations. One patient with four additional mutations, including two co-occurrent variants in the DNMT3A gene, mutation in the TP53 gene and U2AF1 gene, transformed from ET to MF and then from MF to AML. After alloHSCT, the patient had a relapse and died. Two patients diagnosed with myelofibrosis transformed into AML, both patients had additional mutations.

Distribution of additional mutations

For further analysis, due to the NGS test being performed at different time points, patients were divided into two groups. The first group included patients who underwent NGS during initial diagnosis or before the implementation of cytoreductive treatment (n = 31). The second group included those who underwent the test during later genetic diagnostics, after the introduction of methods detecting mutations in the CALR gene or during disease progression (n = 11).

For Group 1 (n = 31), follow-up data were available for 28 patients, ranging from 5 to 195 months with a median follow-up time of 49 months. The median CALR mutation allele burden in the first group was 36.2% (9.3–85.5%). To estimate allele burden, we used VAF detected by NGS. A moderate correlation was found between age at diagnosis and CALR allele burden. There was no difference between CALR allele burden and sex or diagnosis of ET or MF. Information on splenomegaly was available in 24 patients. We observed that patients with splenomegaly had a higher allele burden (p = 0.0008).

We observed that high CALR mutation burden correlated positively with leukocyte count, whereas it correlated negatively with hemoglobin level and weakly negatively with platelet count.

Moreover, we were able to obtain information on disease transformation for 25 patients. In 5 patients, transformation to MF or AML occurred. The CALR mutation burden was significantly higher among patients who experienced disease transformation (p = 0.001). Patients who died shortly after diagnosis, median 15 months (range 5–56), had significantly higher allele burden than patients who survived to the end of follow-up (p = 0.0002).

In this group, additional mutations had 45.2% of patients. We observed that the CALR allele burden was slightly higher in patients with additional mutations, but without statistical significance (p = 0.096). We observed that co-occurring mutations were more common in older patients (p = 0.002).

The study group size was too small to determine whether the presence of additional mutations influenced disease transformation to MF or AML. Although we observed that among 5 patients who had disease transformation, four patients had additional mutations. Two of them had mutations from the HRM group.

Only one patient had thromboembolic or hemorrhagic episodes before diagnosis, but we were unable to obtain information on the type of episodes. He had an additional mutation in the TET2 gene. One patient had an ischemic stroke after diagnosis and had no additional mutations. For three patients, the medical history did not contain information about thromboembolic or hemorrhagic episodes. The remaining patients did not report any episodes.

The second group was much smaller and consisted of 11 patients diagnosed only with ET. The NGS panel testing in this group was conducted several years after diagnosis. Median time 11 years (range, 4–24). For Group 2, follow-up ranged from 72 to 315 months with a median follow-up time of 208 months. The median CALR mutation allele burden was 40.8% (30.5–72%). There was no difference between CALR allele burden and sex. We also observed that CALR allele burden was not associated with the occurrence of additional mutations (p = 0.202).

In six patients, transformation to MF or AML occurred. In this group, the CALR mutation burden was not significantly higher among patients who experienced disease transformation (p = 0.230). It should be emphasized that the NGS test in this group was performed several years after the diagnosis was made. Along with the progression of the disease, the burden allele could also increase.

In this group, additional mutations were observed in 54.55% of patients. Among this group were six patients who transformed to MF or AML. Five of them had additional mutations. Four patients had mutations from the HRM group. Only one patient had thromboembolic or hemorrhagic episodes after diagnosis. The patient had an ischemic stroke after aortic valve replacement. This patient had a CALR type II mutation and two additional mutations in the MPL and SH2B3 genes. Both genes had variants of uncertain significance.

One of the limitations of our study is that the NGS testing was done at different time points. In some patients, NGS was performed at the beginning of the diagnosis; in others, it was performed during complementary molecular tests to search for mutations in the CALR gene after the method assessing the presence of mutations in this gene was introduced. However, the former limitation is somewhat offset by the fact that the majority of patients with CALR mutations in our cohort had NGS testing carried out before the first intervention/treatment. Among the 42 patients, 10 (23,8%) died, 32 (76,2%) survived, and 8 (19%) were out of contact for follow-up.