Introduction

Acute promyelocytic leukaemia (APL) is the most curable subtype of adult acute myeloid leukaemia (AML) with a complete remission (CR) rate greater than 90% [1,2,3,4,5,6,7,8,9,10,11,12]. This has been achievable due to improvements in the diagnostic tools and supportive care, as well as the introduction of all-trans-retinoic acid (ATRA) [1,2,3,4,5,6,7,8,9] and arsenic trioxide (ATO) into the treatment [10,11,12]. Nowadays, a part of solving early death, another important challenge in the treatment of APL patients is to reduce the long-term events, such as second neoplasms (s-NPLs).

Though the reported incidence of s-NPLs after APL therapy is relatively low, their outcome seems very poor [13,14,15,16,17,18,19,20,21,22,23]. However, there are only a few studies analysing the risk factors for the development of s-NPLs after APL therapy. Among chemotherapy-based regimens, topoisomerase II inhibitors and alkylating drugs are proved to be associated with s-MDS/AML development [24, 25], but these drugs are not used in APL. It is supposed that chemotherapy-free regimens, based on ATRA plus ATO, could reduce the number of s-MDS/AML in APL survivals [20,21,22]. On the other hand, there is a close correlation between long-term environmental arsenic exposure and an increase in solid tumours incidence [26, 27]. Of note, there is still scarce information in the literature concerning s-NPLs after chemotherapy-free APL regimens [20,21,22].

There are two main s-NPLs categories in APL patients: solid tumours and second myelodysplastic syndrome (s-MDS) or acute myeloid leukemia (s-AML). Previous multicentre study by the PETHEMA (Programa para el Tratamiento de Hemopatías Malignas) group characterized patients with s-MDS/AML developed after chemotherapy-based (AIDA) APL protocols. There are still few data concerning the incidence and outcomes of solid tumours after APL therapy [19,20,21,22,23].

This retrospective multicentre study aims to analyse the incidence, risk factors and clinical outcomes of s-NPLs in de novo APL patients treated with 5 consecutive PETHEMA protocols.

Methods

We performed a retrospective analysis on 2670 unselected newly diagnosed APL patients, who were treated according to PETHEMA “chemotherapy based” (APL1996, 1999, 2005, 2012 and 2017 for high risk) and “chemotherapy free” regimen (2017 for low/intermediate risk) between 11.1996 and 11.2021. Treatment protocols are summarized in Table 1 [4, 5, 28, 29] and study selection in Fig. 1. The study was conducted in accordance with the Declaration of Helsinki. Informed written consent to follow the treatment according to PETHEMA protocol was obtained.

Table 1 Therapy of APL in the consecutive PETHEMA trials
Full size table
Fig. 1
figure 1

Patient’s selection for the study. Diagnosis of APL was confirmed by the presence of t(15;17) in conventional cytogenetic analysis, and/or detection of the PML/RARa with fluorescence in situ hybridization (FISH) or standardized reverse transcription-polymerase chain reaction (RT-PCR) analysis with a sensitivity of 103-104. Only de novo APL patients who achieved complete remission (CR) and who completed the three consolidation cycles were enrolled into the analysis. Patients with APL second to prior neoplasm, as they received chemo and/or radiotherapy before diagnosis of APL, were excluded from the analysis. Patients were treated in Spain, the Netherlands, Argentina, Uruguay, Belgium, Czech Republic, Portugal (PETHEMA registry) and Poland (PALG, Polish Adult Leukemia Group registry); n: number, sNPL: second neoplasm

Full size image

Definitions of CR, molecular remission and persistence, hematological and molecular relapse have been reported previously [30, 31]. Diagnosis of s-MDS/AML was made according to the WHO criteria [32]. APL relapse was ruled out by cytomorphological and cytogenetic/molecular assessments [30, 31]. The diagnosis of other hematological malignancies and solid tumours was confirmed by histopathological analysis of the excised tissues.

The last patient follow-up was updated on December 2022. The following data was collected at diagnosis: age, gender, ECOG (Eastern Cooperative Oncology Group) performance status score, platelet and leucocyte (WBC) counts, Sanz risk score, fever and coagulopathy, FLT3 (FMS-like tyrosine kinase 3) — internal tandem duplication (ITD), karyotype: t(15;17) vs complex karyotype. Complex karyotype was defined as karyotype with minimum 3 independent abnormalities [33]. Survival and cumulative incidence curves were determined using non-parametric methods using Kaplan-Meier estimators [34]. Differences between curves are done by log-rank test and p-value is given [35]. The probability of developing s-NPL was also estimated by cumulative incidence and the estimated variance was calculated using Aalen–Johansen estimators [36, 37]. The cumulative incidence of n-NPL was calculated from the date of CR. For the cumulative incidence analysis, death in CR and relapse of APL were considered competing causes of failure. Multivariate analysis was performed using the Cox proportional hazards model [38]. Overall survival (OS) was calculated from the date of development of s-NPL. The chi-2 test, with Yates’ correction if necessary, and Mann-Whitney U test and t test were used to analyse differences in the distribution between patient subsets of categorical and continuous nonparametric and parametric variables, respectively. Median lifetimes were determined taking into account censored data. Even if median survival has been reached in a group, it was not possible to calculate complete confidence intervals for those median values. The odds ratio was calculated by median-unbiased estimation. Confidence intervals are calculated using exact methods (mid-p). Characteristics selected for inclusion in the multivariate analysis were those for which there was a significant association in univariate analysis (p .05). All p values reported are two-sided. All tests were performed in R version 4.2.1.

Results

Patient selection

Between November 1996 and November 2021, 3364 APL patients were registered and were intended to be treated according to PETHEMA protocols in Spain, the Netherlands, Argentina, Uruguay, Belgium, Czech Republic, Portugal and Poland. Information on enrolment, eligible patients, lost to follow-up, and exclusion from analysis is shown in a flow diagram (Fig. 1). Of them, there were 2670 de novo APL patients who achieved complete remission and completed induction and all consolidation cycles. Patients included into the study, were treated according to subsequent PETHEMA protocols: LPA1996 (n=169), LPA1999 (n=639), LPA2005 (n=1135), LPA2012 (n=469), LPA2017 (n=258). Main clinical and analytical characteristics of patients included into analysis are presented in a Table 2.

Table 2 Main characteristics of analyzed APL patients according to development of different type of s-NPLs
Full size table

Incidence of second NPLs

Median follow-up of the series was 77 months (range, 17 to 158 months) from diagnosis. Out of 2670 APL patients, 118 (4.4%) developed s-NPLs during follow-up with the median latency period (between first CR and diagnosis of s-NPL) of 48.0 months (range: 2.8–231.1): 43.3 (range: 2.8–113.9) for s-MDS/AML and 61.7 (range: 7.1–231.1) for solid tumour (excluding other hematological neoplasms).

The 5-year CI of all s-NPLs was of 4.43% and 10 years of 7.92%. The CI according to protocols (APL1996, 1999, 2005, 2012, 2017) was as follow: 4.33% vs 3.69% vs 4.91% vs 5.14% vs 1.27% (5 years) and 8.94% vs 6.07% vs 9.61% vs 8.83% vs incomplete observation (10 years) (Table 3, Supplementary Fig. 1).

Table 3 Comparison of cumulative incidence rates of s-NPLs according to type of second neoplasm, PETHEMA protocol, patient’s characteristic
Full size table

Among s-NPLs, there were 58 cases of s-MDS/AML, 3 cases of other hematological neoplasms (2 myeloproliferative neoplasms and 1 Hodgkin lymphoma), 57 solid tumours and 1 non-identified neoplasm. CI according to the type of s-NPLs is presented in Table 3, Supplementary Fig. 2.

Information concerning the type of solid tumour was available in 54/57 (94.7%) patients. The most frequent solid tumour was colorectal, followed by lung and breast cancer (Fig. 2A).

Fig. 2
figure 2

Characteristics of second neoplasms. (A) Second solid tumour types. (B) Cytogenetical abnormalities in second MDS/AML

Full size image

Regarding s-MDS/AML, there was information concerning karyotype abnormalities in 34/58 cases (58.6%) and the most frequent cytogenetic abnormalities were complex karyotype, monosomy of 7, and del 7q (Fig. 2B).

Risk factors for development of s-NPLs

In patients developing s-NPLs there was a slight predominance of female sex (57.6%) and the median age at time of diagnosis of APL was higher vs APL without s-APLs: 55 (range from 13 to 84) vs 43 (range from 2 to 87) years. There were no differences in the Sanz risk score distribution (Table 2).

The following risk factors were analysed as possible risk factors for development of all s-NPLs and s-MDS/AML and solid tumour separately: age, gender, ECOG and Sanz risk score, fever and coagulopathy, WBC and platelet counts, karyotype t(15;17) vs t(15:17)+others, FLT3-ITD mutation at diagnosis of APL, chemotherapy-based vs chemotherapy free regimens. Multivariate analysis identified only age ≤ 35 years (hazard ratio = 0.2584; p < 0.0001) as an independent prognostic factor for all s-NPLs (Table 3, Fig. 3B); we found no significant risk factor for development of s-MDS/AML or solid tumour separately probably due to low number of patients.

Fig. 3
figure 3

Cumulative incidence of s-NPLs according to risk factors. (A) Cumulative incidence of s-NPLs according to type of PETHEMA protocol: chemotherapy-based vs chemotherapy free. (B) Cumulative incidence of s-NPLs according to the age (<35 vs >= 35 y.o.) at diagnosis of APL

Full size image

Respect to type of therapy, all cases with s-MDS/AML (n=117) received chemotherapy-based PETHEMA protocols (APL96 – n=14, APL99 – n=35, APL2005- n=69 and APL2012 – n=12). After chemotherapy-free protocol, there was only one case of s-NPL and it was a solid tumour. CI of s-NPLs according to chemotherapy-based vs chemotherapy-free regimens is presented in a Fig. 3A. In multivariate analysis, there were no significant differences of s-NPLs development between chemotherapy-based vs chemotherapy-free regimens (hazard ratio = 1.09; p =0.932).

Outcomes of patients developing s-NPLs

Overall, the 2-year OS from diagnosis of s-NPLs was 40.6%, with a median OS of 11.1 months (Fig. 4A, Table 4). The 2-year OS from diagnosis of s-MDS/AML and from diagnosis of solid tumours was 31.5 % vs 47.4% with median OS of 9.4 months vs 13.7 months respectively (Fig. 4B). In all patients who died, the main cause of death was s-NPLs itself.

Fig. 4
figure 4

Overall survival estimated by Kaplan-Meier method according to the occurrence of s-NPLs. (A) Occurrence of all s-NPLs. (B) Occurrence of s-MDS/AML vs solid tumours

Full size image
Table 4 Frequency of deaths and median follow-up of APL patients after development of s-NPLs (according to PETHEMA protocol)
Full size table

Discussion

Our study shows that s-NPLs, both s-MDS/AML and second solid tumours, represent a relatively frequent and poor prognosis late complication in patients with APL. The only risk factor related with the development of all s-NPLs was older age (> 35 years old at diagnosis of APL). Our data suggest a potential reduction in incidence of s-NPLs after chemotherapy-free regimens. Nevertheless, longer follow-up is needed to affirm that chemotherapy-free protocols could reduce the risk of developing s-NPLs. As APL is currently the most curable subtype of AML, all efforts should be focused on reducing long-term toxicities, including s-NPLs.

Generally, the occurrence of s-NPL in survivors of adult cancer has risen from 9 to 19% of all neoplasms diagnosed in 1975–1979 and in 2005–2009 respectively [39]. In case of APL patients, data respect s-NPL are very rare. According to previous publications, the crude incidence of s-MDS/AML ranges between 0.97 and 6.5% [17, 18], and according to a previous PETHEMA analysis the 6-year CI of s-MDS/AML was 2.2% [14]. This result was confirmed in the present publication (a 5-year CI of s-MDS/AML was of 2.8%), while a 10-year CI was almost double (4.03%). The crude incidence of second solid tumours ranges between 1.52 and 18%, but the number of reports is even more limited [14, 20, 21]. In the present study, with longer follow-up, we enlarge the prior PETHEMA series and we analyse the CI of solid tumours and other hematological malignancies as well, resulting in an overall s-NPLs CI of 4.4% at 5-year CI and 7.9% at 10 years. Of note, we show that with longer follow-up, the CI of s-NPLs will clearly increase, in particular for solid tumours were 10-year CI was more than double that of 5-year CI (3.7% vs 1.5%). Unfortunately, we could not compare our series with an age-matched cohort of non-APL populations, but give the younger median age of APL patients (between 40 and 45 years old), our findings clearly suggest that developing and being treated of APL is leading to a higher risk of s-NPLs.

Like in previous publications [13,14,15,16,17,18,19,20,21,22,23], we corroborate that prognosis of s-NPLs is very poor (40.6% of 2-year OS, with a median OS of 11.1 months) and even worse for s-MDS/AML cases. In case of s-MDS/AML, it is certainly related to the presence of adverse cytogenetic abnormalities observed at time of diagnosis of second neoplasm. Moreover, it could be related to cumulative toxicity of previous therapies, including that for APL. Importantly, in our series, the main cause of death after development of s-NPLs was always related to resistance to therapy, progression of second neoplasm or side effects related with the treatment of s-NPLs.

Risk factors for s-NPLs development are still not clearly defined. In a present study, we observed that s-MDS/AML cases occurred only among APL patients treated with chemotherapy-based regimens. According to previous PETHEMA analyses, the incidence of s-MDS/AML between patients treated with APL96 vs anthracycline-reinforced APL99 protocol was similar. Intermediate risk APL patients obtain higher dose of anthracycline that could be a risk for s-MDS/AML development [14]. Nevertheless, we have not observed correlation between Sanz risk score and all s-NPLs or s-MDS/AML development. The impact of maintenance therapy with methotrexate and mercaptopurine is still not clear [14]; however, all patients (except one treated with APL2017 protocol) who developed s-NPLs, received maintenance therapy. The diagnosis of s-MDS/AML with abnormal karyotype (complex karyotype, abnormalities of chromosome 5 or 7) suggests a close relationship with prior cytotoxic agents [14,15,16,17,18, 24, 25]. Complex karyotype and chromosome 5 and 7 abnormalities are associated with use of alkylating agents. In APL therapy topoisomerase II inhibitors (anthracyclines) are used, which have been shown to be associated with the 11q23 rearrangement (KMT2A, previously MLL) [24, 25]. In this context, the role of topoisomerase II inhibitors in leukemogenesis of s-MDS/AML in APL patients remains unclear. Moreover, at the time of diagnosis of APL, additional chromosomal abnormalities (trisomy 8 and abnormalities in chromosome 7 among others) are quite frequent (28–30%) [33, 40,41,42]. Probably, due to low number of patients with additional karyotype, the present study was not able to determine the role of additional chromosomal abnormalities on the development of s-NPL nor s-MDS/AML in APL patients after achieving of CR.

The age more than 35 years old was suggested as an independent risk factor for development of s-MDS/AML [14]. Based on the present analysis, we confirmed that age of 35 or more is an independent risk factor for development for all s-NPLs but not for s-AML/MDS and solid tumours separately.

Data concerning the incidence of s-NPLs after chemotherapy-free regimens are still scarce, but they seem to be equal [20,21,22] or less frequent than after chemotherapy-based regimens. There is a close correlation between exposure to inorganic arsenic and an increase in lung, kidney, liver, bladder and skin cancers incidence [26, 27]. We were unable to show reduction of s-NPLs after chemotherapy-free regimens compared to AIDA-based protocols probably due to: (1) the limited number of patients treated with chemotherapy-free regimens; and (2) the shorter follow-up of these patients, which is very relevant taking into account that median latency to occurrence of solid tumours overlaps with longer follow-up of ATO+ATRA patients in our series.

The incidental occurrence of s-NPLs in APL patients cannot be ruled out. Many risk factors such as increased body mass index (BMI), diet, cigarette smoking, low physical activity could be responsible for the development of s-NPLs in the general population. Interestingly, the majority (59%) of APL patients are overweight or obese (BMI ≥25) [43].

Our study has several limitations related to retrospective design; there are missing data concerning the molecular and cytogenetical analysis at time of diagnosis of APL and of s-NPLs. The number of patients with s-NPL and in particular the group of patients treated with ATRA plus ATO is very small and with short follow-up period. Despite these limitations, this study was performed on a large group of APL patients treated homogenously.

To conclude, s-NPLs are relatively frequent in APL patients. We report higher incidence than previous reports as patients had substantial follow-up. The development of s-NPLs was associated with dismal outcomes, highlighting the need of preventive strategies, long-term monitoring and high suspicion after APL therapy. Further studies are needed in order to assess the potential reduction of the risk of s-NPLs among patients treated with chemotherapy-free regimens.