Abstract
High incidence and prognostic relevance of thromboembolic disease in patients with ALK-rearranged NSCLCs http://ow.ly/DEZr30j6kC8
To the Editor:
Thromboembolic disease is fairly common in patients with lung cancer [1–3]. This incidence seems to be higher in patients with lung adenocarcinomas [4], with approximately 15% of those with advanced stage disease developing venous thromboembolisms (VTE) during the whole course of their disease [5–7]. Pulmonary adenocarcinomas are a heterogeneous group of diseases that can be stratified according to the presence of major oncogenic driver alterations. Anaplastic lymphoma kinase (ALK) rearrangements are detected in approximately 4% of these cases [8]. Isolated reports have suggested that patients bearing ALK-rearranged tumours might have a higher than expected incidence of thromboembolisms [9, 10]. In the present study, we have analysed the incidence, predictors and prognostic significance of thromboembolic events in a large, multi-institutional and homogeneous cohort of advanced stage patients with ALK-rearranged lung cancers from Spain and Portugal. Our primary objective was to estimate the incidence of thromboembolic events and their association with overall survival in these patients.
A centralised institutional ethics committee approval at the 12 de Octubre University Hospital valid for all Spanish centres, and an institutional ethics committee approval at the Portuguese Institute of Oncology of Porto, were obtained before the study was initiated. We retrospectively selected all consecutive patients diagnosed with advanced stage (stages III and IV) ALK fusion positive non-small cell lung cancers (NSCLCs) between January 2012 and December 2016. Data were contributed by 29 Medical Centres from Spain and one from Portugal. ALK positivity was determined according to local standard protocols in each institution. We excluded patients with neuroendocrine tumours and patients on therapeutic doses of anticoagulants prior to advanced stage cancer diagnosis. We defined a thromboembolic event as any venous or arterial thromboembolism, documented by imaging studies, that occurred at the time or after advanced stage cancer diagnosis. In addition to thromboembolic events, collected during the whole patients' follow-up period, we collected baseline information (within 1 month of advanced stage cancer diagnosis) of several clinical and analytical variables of interest.
We included 241 ALK-rearranged NSCLCs in this study. The median age was 56 years (range 17–84 years). Half of the patients were never smokers (52%), and most had stage IV pulmonary adenocarcinomas (n=204, 85%). Baseline brain and liver metastasis were detected in 22% and 25% of the patients, respectively. 17 patients (7%) and 185 patients (77%) had high and intermediate Khorana risk scores (KRS) [11] respectively. The median follow-up of our study population was 19 months (range 0–59 months), and 127 (53%) of the patients died. The median follow-up of alive patients was 30 months (range 4–49 months). The estimated median overall survival was 26 months for the entire series, and 31 months for those patients that received ALK tyrosine kinase inhibitors (TKIs) (n=207, 86%).
73 patients (30%) experienced thromboembolic complications. 71 out of 73 patients (97%) had VTE, with pulmonary embolism as the most frequent location (n=30; 41%). Four patients (5%) developed arterial thromboembolisms, two of them with concurrent VTE. Thromboembolic events occurred within the first 6 months from cancer diagnosis in 54 patients (74%), of whom 24 (33%) were detected within the first month from diagnosis. 22 patients (30%) were deemed to have incidental (asymptomatic or clinically unsuspected) thromboembolisms. 12 patients (16%) developed recurrent thrombosis (seven while receiving appropriate therapeutic anticoagulation). The mortality directly attributed to thromboembolic disease in this series was 2% (n=5).
To study the association between baseline characteristics and thromboembolic disease, we first conducted a univariate competing risk regression analysis [12] including gender, age, body mass index, performance status, smoking history, histology, tumour stage, central nervous system metastasis, liver metastasis, number of metastatic sites, haemoglobin, leukocyte, platelet and albumin counts, INR range and KRS. In this analysis, the presence of baseline liver metastases, leucocyte counts >11 000 cells per mm3, three or more metastatic organ involvement and a high KRS (HR 2.82, CI 95% 1.48–5.37; p=0.002) were significantly associated with increased risk of thromboembolic disease. In the multivariate model (in which we did not include KRS because it already incorporates leukocyte counts), only liver metastases and leukocytosis remained as independent predictors of thromboembolic disease, with corresponding HR of 1.85 (CI 95% 1.09–3.15; p=0.021) and HR of 2.34 (CI 95% 1.43–3.82; p=0.001), respectively.
Patients that experienced thromboembolic events at any time point had shorter median overall survival (20 months) than patients without thrombosis (36 months) (p=0.035) (figure 1a). Patients with thromboembolic events at baseline (n=24) had a median overall survival of 15 months (figure 1b). Patients that experienced recurrent thromboembolisms (n=12) had a median overall survival of 10 months, without statistically significant differences when compared with those without recurrent thromboembolism (n=61; 21 months) (p=0.369). Considering the development of thrombotic events as a time-dependent covariate, thromboembolic disease was significantly associated with an increased risk of death (HR 2.22, CI 95% 1.53–3.21; p=0.000). We conducted a multivariate Cox regression analysis including smoking history, comorbidities, leukocytosis, liver metastasis, performance status and ALK TKI treatment as fixed covariates, and thromboembolic disease as a time-varying covariate. In this analysis, thromboembolic disease remained as an independent predictor of overall survival with HR 1.70 (CI 95% 1.10–2.62; p=0.016) (figure 1c). The presence of thromboembolic disease at baseline (n=24) was associated with a numerically nonsignificant increased risk of death (HR 1.67, CI 95% 0.96–2.91; p=0.068).
In the present study, we show that thromboembolic disease is a frequent complication in patients with advanced-stage ALK-positive NSCLCs (30%), particularly in the presence of baseline liver metastasis or leukocytosis (∼50%), and it is associated with a lower overall survival in this particular subtype of lung cancer. This study was not designed to compare the incidence of thromboembolic events between different molecular subtypes of lung cancers. However, the high incidence observed here is consistent with what has been reported in smaller ALK-positive cohorts [9, 10], and exceeds the observed in other subtypes of lung cancers [2, 3, 5–7]. We found that thromboembolic disease was an independent prognostic factor in advanced ALK-positive NSCLCs. Although the mortality rate of thromboembolic complications might have been underestimated in this study [13], the differences in survival do not seem to be directly related with the fatality of thromboembolic events themselves (2% of mortality attributed to thromboembolic disease in this cohort). In the present study, both leukocytosis and liver metastasis significantly predicted both the occurrence of thrombotic complications and reduced overall survival (figure 1c). A relatively low proportion of patients had progressive disease at the time of thromboembolic complications (22%), but we cannot definitely exclude that the association of thromboembolic disease with lower overall survival was confounded by subsequent development of liver metastasis and/or leukocytosis in some patients, as we only had baseline information for these variables.
Our findings have relevant implications for clinical practice. First, physicians should actively examine for the presence of signs or symptoms of thrombosis at diagnosis and during follow-up of these patients, and accordingly rule out the existence of thromboembolic disease with appropriate imaging studies. Additional surveillance strategies in high risk patients, including those with leukocytosis or liver metastasis, would need to be eventually considered. Second, these results justify investigation, in the context of a randomised trial, of the clinical benefit, in terms of thromboembolisms risk reduction and potentially in overall survival improvement, of adding prophylactic doses of anticoagulants to standard anticancer treatments in these patients. And ultimately, with median overall survival estimates exceeding 4 to 5 years with the incorporation of novel-generation ALK TKIs into the clinic [14, 15], thromboembolic disease will be one of the main cancer-related complication during the whole course of the disease in these patients.
Footnotes
Author contributions: J. Zugazagoitia, A. Manzano, L. Paz-Ares, S. Ponce-Aix and A.J. Muñoz conceived and designed the study. All authors were involved in data collection. D. Lora, together with J. Zugazagoitia and A. Manzano, did the statistical analysis. J. Zugazagoitia, A. Manzano, D. Lora and L. Paz-Ares contributed to data interpretation. J. Zugazagoitia and A. Manzano, together with L. Paz-Ares, wrote the original draft of the manuscript. All authors participated in manuscript writing and editing, and approved the final version of the article.
Conflict of interest: M. Biosca has received personal fees from Leo-Pharma and Sanofi, outside the submitted work.
Conflict of interest: L. Paz-Ares has received honoraria from Roche, Lilly, Bristol Meyers Squibb, Merck & Sharp D., Boehringer Ing., Pfizer, AstraZeneca, Amgem and Novartis, outside the submitted work.
Conflict of interest: A. Manzano has received personal fees from Pharmamar, Roche and AstraZeneca, outside the submitted work.
Support statement: The Thrombosis Working Group of the Spanish Society of Medical Oncology supported the study and contributed to the distribution of the protocol to participating centres. M. Biosca, C. Font, I. García Escobar, C. Salvador-Coloma, M. Sánchez Cánovas, B. Obispo, J. Calzas, A.J. Muñoz and A. Manzano are members of the Thrombosis Working Group of the Spanish Society of Medical Oncology. J. Zugazagoitia was funded by Instituto de Salud Carlos III (Rio Hortega, CM15/00196). L. Paz-Ares was funded by ISCIII: PI1401964, PIE15/00076, RTICC (R12/0036/0028) and CIBERONC (C16/12/00442), co-funded by FEDER from Regional Development European Funds (European Union).
- Received August 8, 2017.
- Accepted February 24, 2018.
- Copyright ©ERS 2018