Incidence and Mortality of Mycosis Fungoides and Sezary Syndrome: A Nationwide Registry Study in Finland

Abstract

This study investigated the incidence, mortality, and causes of death in Finnish patients with mycosis fungoides (MF) or Sezary syndrome (SS). The study population consisted of 450 patients with MF and 23 with SS and their 1:1 matched controls. Data were collected from nationwide Finnish healthcare registries. The average incidence of MF was 0.38 and of SS 0.02 per 100,000 persons in 2007–2019 and was highest among older adults. Patients with SS had a low 5-year overall survival of 33.5% vs 64.5% in their controls (p = 0.0046), whereas 62.3% of patients with MF were alive 10 years after diagnosis vs 71.4% of their controls (p = 0.0011). SS diagnosis (hazard ratio [HR] 11.85 [5.08, 27.65]; p < 0.001) and higher age (HR 1.07 [1.01, 1.12]; p = 0.01) were risk factors of poorer disease-specific survival. Moreover, patients with MF had a 2.38-fold [1.07, 5.27] risk of thyroid diseases (p = 0.033) and over 1.5-fold [1.17, 2.07] risk of cardiovascular diseases (p = 0.002) compared with controls. In conclusion, in the Finnish population, patients with MF or SS had a higher incidence and mortality and more late comorbidities than controls. More awareness, earlier diagnostic tools, and more effective treatment for MF/SS are needed.

SIGNIFICANCE

This study aimed to clarify the incidence and mortality of 450 patients with mycosis fungoides and 23 patients with Sezary syndrome in Finland. We found an increased trend in incidence of mycosis fungoides. In addition, we observed that older age at diagnosis and, especially, the diagnosis of Sezary syndrome compared with mycosis fungoides were associated with a higher risk of mycosis fungoides or Sezary syndrome disease-specific mortality. Moreover, patients with mycosis fungoides or Sezary syndrome demonstrated higher cardiovascular disease-related morbidity and mortality than their matched controls. Awareness of mycosis fungoides or Sezary syndrome, earlier diagnostic tools, and more effective treatments are needed to improve patient outcomes.

Cutaneous lymphomas account for 20% of extranodal lymphomas and predominantly arise from T cells (1). Mycosis fungoides (MF) accounts for almost half of all cutaneous lymphomas and 60% of cutaneous T-cell lymphomas (CTCLs). In contrast, Sezary syndrome (SS) accounts only for 2% of CTCLs (1, 2). According to SEER registry data, the annual incidence of MF is 5.42 and that of SS 0.21 per 1,000,000 persons in the USA (3, 4) with a male predominance (2, 5). For unknown reasons, the incidence of these rare and heterogeneous lymphoma entities has been increasing in several countries (3).

In contrast to most systemic T-cell lymphomas with adverse outcomes and high mortality, an indolent clinical course is typical for most CTCLs. Patients diagnosed with early stages of MF rarely die because of lymphoma; however, the higher stage of the disease also increases the risk of lymphoma-associated death (2, 4). Moreover, several studies have reported that patients with MF have increased mortality associated with comorbidities such as diabetes mellitus, atherosclerosis, and other cancers possibly caused by the hyperinflammatory nature of MF (6).

Scarce data exist on the incidence and detailed causes of death in patients with MF or SS. Therefore, this nationwide registry study aimed to describe the incidence patterns of MF and SS in Finland and clarify lymphoma-related and other mortality among these patients.

MATERIALS AND METHODS

This retrospective nationwide registry study included 450 Finnish patients with MF and 23 patients with SS and their 1:1 matched controls and used data generated during routine clinical practice in primary and specialty care available in the Finnish electronic healthcare registries. Four nationwide data controllers were utilized for data collection: the Finnish Cancer Registry (FCR), National Institute of Health and Welfare (THL), Statistics Finland and Digital, and Population Data Services Agency (DPA). The data covered the entire population of Finland, accounting for approximately 5.6 million people in 2019. The unique Finnish identification number ensured matching between registries.

Patients with an initial record of MF or SS in FCR between 1 January 2007 and 31 December 2019 were included in the analysis cohort (n = 473). The inclusion was based on ICD-O-3 morphological codes 9700 and 9701. An age-, sex-, and region-matched control patient cohort in a 1:1 ratio was then identified using DPA records. The criteria for eligible control participants included no lymphoma diagnosis (C81-C85) in FCR records and status alive at the index of the corresponding case.

Data analysis

Demographic characteristics were assessed at the diagnosis date (index). The Charlson comorbidity index (CCI) (7) was calculated based on co-diagnoses recorded during 3 years until 6 months prior to the index. Lymphoma diagnoses (ICD-10: C81-C85) were excluded before CCI calculation for comparable scores with controls. The proportion of patients treated with radiotherapy (RT) was estimated with a competing risk model with a competing risk of death.

Summary statistics were presented as median and interquartile range for continuous variables and as numbers and percentages for categorical variables. The differences between cases and controls were analysed using the Wilcoxon signed-rank test for continuous and McNemar’s test for categorical variables. In order to retain anonymity of the patients with SS, some analyses were made of the whole population including both MF and SS patients.

The follow-up (FU) for patients started on the date of MF or SS diagnosis. The control participants were followed from the predetermined index of their corresponding cases. The end of FU was defined as death or end of the study (31 December 2019), whichever occurred first. Overall survival (OS) was analysed using the Kaplan–Meier method, and patients alive at the end of FU were censored. For the survival estimates, 95% confidence intervals were provided. Differences between the survival of cases and controls were analysed using a stratified log-rank test accounting for the matched setup.

The annual incidence of MF and SS per 100,000 was calculated based on the annual population sizes in Finland. The trends in incidences were analysed using the Kendall rank correlation coefficient (Kendall’s tau) and analysed using the Mann–Kendall trend test. Age-standardized incidence was scaled by the global population age distribution by the World Health Organization (8). Incidence by age group according to sex was also calculated.

The association between demographic, socioeconomic, and clinical features with OS was analysed using a Cox proportional hazards model. Hazard ratios (HRs) for age and diagnosis year are reported per year. The proportional hazards assumptions were tested.

The relative risks of long-term comorbidities were compared with controls using Cox models, in which the first record of comorbidity was treated as an event, and the end of FU and death were censoring events. Baseline comorbidities were assessed from 3 years before the index until 3 months after the index, and participants with baseline comorbidities were excluded from the analysis. A clustering variable for matched case-control pairs maintained the matched analysis. The studied comorbidities included cardiovascular disease (CVDs) (I10–I25, I34–I37, I42–I43, I46, I50), thyroid disease (E00–E07), diabetes (E10–E14), lung disease (J00–J06, J09, J10–J18, J20–J22, J39–J47, J60–J70, J80–J86, J90–J99), myeloid malignancies (C92, D46, D47), skin cancer (C43–C44), solid malignancies excluding skin cancer (C00–C39, C40–C42, C47–C47, C50–C79, C80), and mental health problems (F00–F48, F50–F59, R53).

The cause-specific cumulative mortality to MF or SS (C84.0–C84.1), other solid malignancies (C00–C80), myeloid malignancies (C92, D46, D47), CVDs (I10–I25, I34–I37, I42–I43, I46, I50), infections (A*, B*, J00–J22, J85–J86, N10, N30), and other causes (any other cause of death) were estimated.

All analyses were performed using R 4.0.3 (R Foundation for Statistical Computing, Vienna, Austria). All tests were two-sided unless otherwise stated, and statistical significance was set at 0.05. All results and p-values are descriptive, and no multiple testing correction was applied.

RESULTS

Study population

A total of 450 of the patients included were diagnosed with MF and 23 with SS. The baseline demographics of the patients and controls are presented in Table I and Table SI. The median age of the population in MF patients was 66 years (55.67–76.66) and 72.2 years (65.16–76.08) in SS patients, with male predominance in both groups (59.3% and 56.5%). The patient and control groups were well balanced by patient demographics. Only the marital status differed between the control and patient groups (p = 0.042) at baseline. RT was administered only to study patients: 6.1% [4.27, 8.55] of patients were treated with RT within 1 year after diagnosis. The median observed FU time was 4.2 (1.89–7.04) years for MF patients and 4.60 (2.05–7.79) for their controls (p = 0.013) and 1.95 (0.59–5.45) for SS patients and 5.72 (2.94–7.96) for their controls (p = 0.005).

Table I.

Demographics of 450 patients with mycosis fungoides (MF) and 23 patients with Sézary syndrome (SS) and their matched controls

Factor	Level	MF patients	Controls	p-value	SS patients	Controls	p-value
Age (median, IQR)		65.99 [55.67,76.66]	65.99 [55.67,76.66]		72.24 [65.16,76.08]	72.24 [65.16,76.08]
Gender, n (%)	Female	183 (40.7)	183 (40.7)		10 (43.5)	10 (43.5)
	Male	267 (59.3)	267 (59.3)		13 (56.5)	13 (56.5)
CCI, n (%)	0	385 (85.6)	369 (82.0)	0.375	21 (91.3)	20 (87)	NaN
	1–2	57 (12.7)	67 (14.9)		< 5	< 5
	3+	8 (1.8)	14 (3.1)		< 5	< 5
Follow-up time, years		4.20 [1.89,7.04]	4.60 [2.05,7.79]	0.013	1.95 [0.59,5.45]	5.72 [2.94,7.96]	0.005

Incidence

The annual age-standardized incidences of MF and SS from 2007 to 2019 are presented in Fig. 1. During the study period, the average incidence of MF was 0.38 and of SS 0.02 per 100,000 persons in 2019 (Fig. 1A). Age-standardized increase in the annual incidence of MF was seen especially in females (Kendall tau = 2.5, p = 0.01) (Fig. 1B). Higher insidences were observed in older age groups (Fig. 1C, Table SII): the incidences at the age groups of 30–34 years, 50–54 years, and 80–84 years were 0.2, 0.66, and 2.98 per 100,000 persons, respectively.

Fig. 1.

(A) Age-standardized annual incidence of MF and SS, (B) age-standardized incidence of MF and SS per sex, and (C) incidence according to age groups of MF and SS in Finland.

Survival and causes of death in patients with mycosis fungoides or Sezary syndrome

The 5-year survival rate for patients with MF was 79.5% (74.8–83.8) vs 87.3% (83.2–90.5) in controls. A total of approximately 62.3% (54.9–68.7) of patients with MF were alive 10 years after diagnosis compared with 71.4% (64.1–77.4) in controls (p = 0.0011). The 5-year OS for patients with SS was 33.5% (15.3–52.8) vs 64.5% (41.3–80.4) in controls. The median OS in patients with SS was significantly lower than that in controls (median OS 23.4 months vs NR, p = 0.0046) (Fig. S1).

In the cause of death analysis with pooled MF or SS, lymphoma caused an estimated 13% of deaths during FU, whereas 6% died of other cancers and 12% of CVDs. In controls, lymphoma was an exclusion criterion; hence, no lymphoma deaths were recorded. An estimated 7% of controls died of other cancers and 7% of CVDs. The cumulative mortality (%) of patients with MF or SS and their controls by cause of death is shown in Fig. 2 and Table II.

Fig. 2.

Mortality rate and causes of deaths among patients with (A) MF and (B) SS and their controls.

Table II.

Cumulative proportions of patients with mycosis fungoides (MF) or Sézary syndrome (SS) and their controls alive and specific reasons for death at various timepoints

Time from diagnosis	Lymphoma type	Group	Alive, %	Died from lymphoma, %	Died from other cancer, %	Died from CVDs, %
1 year	MF	Patients Control	97 97	1	1 0.7	2 0.5
	SS	Patients Controls	65 91	22	0 4	4 0
5 years	MF	Patients Controls	80 67	4	4 3	6 5
	SS	Patients Controls	34 65	40	9 22	9 4
10 years	MF	Patients Controls	62 71	7	6 5	11 8
	SS	Patients Controls	10 53	45	9 34	9 4
13 years	MF	Patients Controls	59 68	7	6 5	13 8
	SS	Patients Controls

CVDs: cardiovascular diseases.

Risk factors for mortality

Increased risk of overall mortality was associated with advanced age (HR = 1.08 per year [1.05, 1.12]; p = 0.001) and higher CCI level (HR = 3.12 [1.93, 5.02]; p < 0.001). In addition, the diagnosis of SS was associated with a 4.69-fold risk for overall mortality (p < 0.001) compared with MF diagnosis. Moreover, tertiary-level education (HR = 0.50 [0.26, 0.97]; p = 0.04) was associated with lower overall mortality risk. Regarding disease-specific mortality, age at diagnosis was associated with a 1.07-fold ([1.01, 1.12]; p = 0.01) risk and diagnosis of SS over 11-fold ([5.08, 27.65]; p < 0.001) risk of mortality (Fig. 3).

Fig. 3.

Risk factors for (A) overall mortality and (B) disease-specific mortality in 473 patients with MF or SS.

Risk for comorbidities after mycosis fungoides diagnosis

During the FU after MF diagnosis or index, patients had a 2.38-fold [1.07, 5.27] risk of thyroid diseases (p = 0.033) and over 1.5-fold [1.17, 2.07] risk of CVDs (p = 0.002) compared with controls (Fig. 4). The risk of haematological malignancies and solid cancer did not significantly differ between the patients and matched controls. Similarly, no difference was observed in the risk of diabetes, lung diseases, or mental health problems.

Fig. 4.

Comorbidity risk for 450 patients with MF.

DISCUSSION

This large nationwide registry study aimed to clarify the incidence and mortality of MF and SS in Finland. We report a trend of increasing incidence of MF during the FU. The incidence of MF increased with age. Patients with SS had a low 5-year survival of 33.5% compared with 64.5% in their controls, whereas 62.3% of the patients with MF were alive 10 years after diagnosis vs 71.4% of their controls. The diagnosis of SS compared with that of MF was associated with an over 11-fold risk and older age at diagnosis with a 1.07-fold risk for MF or SS disease-specific mortality. Moreover, the patients with MF or SS demonstrated higher CVD-related morbidity and mortality than controls. Awareness of MF or SS, earlier diagnostic tools, and more effective treatments are needed to improve patient outcomes.

We observed an increased trend for age-standardized incidence of MF with an average incidence of 0.38 per 100,000 persons during FU and also male predominance among the Finnish population. Similar trends in results of incidence of MF have been reported in Korea (9) and the USA, with an incidence of up to 0.56 per 100,000 persons in 2002–2018 (3). In Norway, the incidence of MF and SS increased from 0.15 to 0.18 per 100,000 person-years from the 1980s to 2000s (10). According to previous data, individuals at highest risk for MF were males and African Americans (11), whereas a Brazilian study described typical patients with MF to be white men in their 50s (12).

The reasons behind these increasing numbers are partly obscure. In the early stages of MF and SS, clinical and histopathological diagnosis is challenging, and improving diagnostics may partly explain these findings. Notably, we did not have data regarding stage distribution, which might have been able to support this hypothesis. However, Keto et al. (13) described an increasing prevalence of MF and SS among the Finnish population, indicating that diagnosis occurs earlier than in previous years, possibly because of increased awareness of these entities. The influence of the changing age structure of Finnish people in terms of the growing proportion of older adults on the increasing incidence of MF and SS needs to be further studied.

One possible reason for increasing incidence is the role of occupational and environmental toxin exposure during MF or SS development; for example, halogenated hydrocarbons and exposure to solar radiation increase the risk of MF (12, 14). The role of solar exposure is further supported by recent findings that describe the mutational spectrum of MF, which is typical of solar damage (12). In a Canadian study of patients with CTCL, of which most were diagnosed with MF, the incidence was highest in the most industrialized regions (15), in concordance with a Tunisian study (16). In the SEER database, the highest incidence seems to be in metropolitan counties (3). In this light, a higher educational status and working conditions may protect from MF, although tertiary-level education was not significantly associated disease-specific mortality in the present study.

Moreover, some drugs, such as hydrochlorothiazide (12,17), and lifestyle habits, such as smoking, have been associated with an increased risk of MF diagnosis (16,18). In addition, genetic predisposition has been suggested as an aetiological factor, although a Danish twin study (19) did not find an increased risk of CTCLs or other malignancies among co-twins of patients with CTCL. Recently, skin microbiome dysbiosis and cutaneous virus have been suggested to be associated with MF, even if the causality of these associations is not well defined (20).

Diagnosis of SS was associated with an elevated risk of overall mortality, in line with several studies showing the poor prognostic effect of SS (21). However, index year was not remarkably associated with improved MF or SS mortality rate in later years in our study, suggesting a minor impact of recent therapeutic advances on patient outcomes. In the present study, the risk for solid cancers and haematological malignancies in patients and matched controls was comparable. The finding could have been affected by the limited number of patients included in addition to limited FU time.

A Finnish study has demonstrated that patients with MF or SS are at increased risk of dying from other cancers (22). Moreover, a more recent American study has reported an increased any-cause mortality in patients with MF or SS compared with that in age-matched controls (23). A large registry study of 6,742 patients with MF based on the USA SEER data reported 7.5% of second cancer events. Of those, the highest standardized incidence ratios were found for haematological malignancies, followed by solid tumours such as melanoma, lung, colon, and renal cancers with sex-specific female breast and male prostate cancers. Women had a higher risk profile in this study (24), in contrast with an earlier study (25). One reason for the increased risk of secondary malignancies may be that at least T-cell lymphomas and Hodgkin’s lymphomas may represent a transformation of underlying CTCL. Another potential explanation for haematological malignancies is a common progenitor cell carrying driver mutations and genotoxic effects of therapies used. However, only a minority of the patients with MF or SS usually receive alkylating agents as part of therapy, and regarding RT only a limited number of patients in the present study received RT before the introduction of new specified guidelines, possibly partly for other malignancies (26). Other potential amplifiers for haematological and solid malignancies are the disease itself and treatment-associated immunosuppression, especially in advanced-stage disease, common genetic predisposition, and exposure to carcinogenic agents like tobacco.

In this paper, we describe an increased risk of CVD during lifetime and CVD-associated death among patients with MF or SS compared with a matched background population. In contrast to our findings, neither a previous Danish study (27) nor a recent publication from the USA (28) found any differences in the risk of CVDs between patients with MF and their matched controls. However, when focusing on the latest period from 1995 to 2010, Lindahl et al. (27) reported an 80% increase in the risk of vascular events within 5 years after MF diagnosis. MF shares features with many other chronic conditions, such as CVDs, and there is probably partly a common aetiology for those diseases. Moreover, systemic retinoids used in the treatment of MF may induce dyslipidaemia as a risk factor for heart diseases (29). In asymptomatic patients with MF or SS, clinical recognition and treatment of CVD risk factors are crucial, as well as controlling biomarkers of thyroid function.

Limitations of our study include the lack of patient-level data that cover stage distribution and detailed treatment characteristics. In addition, the small number of patients with SS limited some disease-specific analyses to retain anonymity of the patients. Moreover, we were not able to perform a retrospective pathology review, which might have produced some inaccuracies. Additionally, a few patients shifted diagnoses from MF to SS during the study, which may have only minimal impact on the results. However, considering that FCR is based on pathology reports, we anticipate that these have probably not caused any major bias. The strength of the current real-world study includes its nationwide, population-based coverage and reliable data extraction from high-quality national registries.

In conclusion, we report an increasing trend in incidence of MF in the Finnish population. Despite the indolent nature of the diseases, approximately 13% of patients still die of lymphoma. Notably, patients also have an increased risk of CVD-related morbidity and mortality compared with controls. This finding implies that clinicians should be highly aware of the possibility of MF or SS, the risks of increased mortality for any reason, and the need for aggressive treatment of CVD risk factors among patients with MF or SS.

Funding Statement

Data collection was funded by Takeda Oy. The authors would like to acknowledge Editage (www.editage.com) for English-language editing.