Phototherapy as a treatment of early-stage mycosis fungoides and predictive factors for disease recurrence: A 17-year retrospective study
How to cite this article: Rattanakaemakorn P, Ploydaeng M, Udompanich S, Thadanipon K, Rutnin S, Rajatanavin N. Phototherapy as a treatment of early-stage mycosis fungoides and predictive factors for disease recurrence: A 17-year retrospective study. Indian J Dermatol Venereol Leprol 2021;XX:1-7.
Mycosis fungoides is the most common form of cutaneous T-cell lymphoma. Narrowband ultraviolet B and psoralen and ultraviolet A are effective treatment options, but studies of their treatment efficacy and disease relapse remain limited.
This study aimed (1) to determine the efficacy of narrowband ultraviolet B and psoralen and ultraviolet A as a treatment for early-stage mycosis fungoides and explore the predictive factors for complete remission and (2) to determine the relapse rate and analyze their predictive factors, including the utility of maintenance therapy.
This was a retrospective cohort study consisting of 61 patients with early-stage mycosis fungoides (IA - IB) treated with narrowband ultraviolet B or psoralen and ultraviolet A as the first-line therapy from January 2002 to December 2018 at the Division of Dermatology, Ramathibodi Hospital, Bangkok, Thailand. Cox regression analysis and Kaplan–Meier survival curve were performed for the main outcomes.
A complete remission was achieved by 57 (93.5%) patients. The median time to remission was 7.80 ± 0.27 months. Types of phototherapy (narrowband ultraviolet B or psoralen and ultraviolet A), age and gender did not associate with time to remission, while the presence of poikiloderma and higher disease stage led to a longer time to remission. The cumulative incidence of relapse was 50.8%. The median time to relapse was 24.78 ± 5.48 months. In patients receiving phototherapy during the maintenance period, a treatment duration longer than six months was associated with a significantly longer relapse-free interval.
Narrow-band-ultraviolet B and psoralen and ultraviolet A are effective treatment options for early-stage mycosis fungoides. Maintenance treatment by phototherapy for at least six months seems to prolong remission.
Plain Language Summary
Mycosis fungoides is the most common form of cutaneous T-cell lymphoma. Phototherapy (narrowband ultraviolet B and psoralen and ultraviolet A) is one of the most common methods for treatment of early-stage mycosis fungoides because of its high efficacy. However, disease relapse after treatment discontinuation is common. Therefore, this 17-year retrospective study was conducted to determine the efficacy of phototherapy as the treatment of early-stage mycosis fungoides, including exploration of the predictive factors for complete remission and disease recurrence, including the utility of maintenance therapy. The results found that 93.5% of patients achieved complete remission. The presence of altered skin color together with thinning of skin and visible blood vessels on the skin and higher disease stage led to a longer time to subside. Maintenance treatment with phototherapy for at least six months was beneficial prolonging remission. In conclusion, phototherapy is an effective treatment option for early-stage mycosis fungoides.
Mycosis fungoides can be described as the most common form of cutaneous T-cell lymphoma.1 A majority of cases are diagnosed at an early stage (95% stage IA – IIA),2 with an incidence rate of 0.36 per 100,000 person-years.1 Mycosis fungoides is diagnosed from the clinicopathological features according to the consensus criteria by the World Health Organization and the European Organization for Research and Treatment of Cancer.3 The treatment aims to clear all lesions to prolong the relapse-free interval and overall survival with least toxicity.1,4
The first-line treatment of early-stage mycosis fungoides is skin-directed therapy which includes the following: topical corticosteroids, topical nitrogen mustard (HN2), topical BCNU (carmustine), psoralen and ultraviolet A, narrowband ultraviolet B, total skin electron beam therapy and superficial X-irradiation.1,5
Phototherapy has been established as one of the most suitable treatments for mycosis fungoides, especially narrowband ultraviolet B and psoralen and ultraviolet A, because of its effectiveness in lesion clearance and minimal side effects. Complete remission rates of narrowband ultraviolet B and psoralen and ultraviolet A were 62.2% and 73.8%, respectively.6
Although phototherapy is an effective treatment option, rapid disease relapse after treatment discontinuation appears to interfere with its efficacy.7 Currently, the information on disease relapse and its predictive factors, including the usefulness of maintenance therapy, is still lacking.
This study aimed (1) to determine the efficacy of narrowband ultraviolet B and psoralen and ultraviolet A as the treatment of early-stage mycosis fungoides and explore the predictive factors for complete remission and (2) to explore the disease recurrence and analyze the predictive factors for disease recurrence, including the utility of maintenance therapy.
This is a 17-year retrospective study of patients with early-stage mycosis fungoides (IA - IB) treated with narrowband ultraviolet B or psoralen and ultraviolet A as the first-line treatment from the division of dermatology at the Ramathibodi Hospital, Bangkok, Thailand between January 2002 and December 2018. Approval from the research ethics committee was obtained with an ID number 12-61-96.
The inclusion criteria included patients diagnosed with early-stage mycosis fungoides by skin biopsy who had started their first narrowband ultraviolet B or psoralen and ultraviolet A course as the first-line treatment and achieved a complete remission by the end of 2018. We excluded all patients who had previous treatments with systemic agents (interferon and retinoids) or other skin-directed agents (topical chemotherapy and electron beam therapy), who had stopped phototherapy before a response was defined (complete, partial or nonresponse), who were lost to follow-up or who had incomplete data on the electronic medical record.
The stage of the disease was classified according to the tumor-node-metastasis-blood (TNMB) classification by the International Society for Cutaneous Lymphoma and Cutaneous Lymphoma Task Force of the European Organization for the Research and Treatment of Cancer.8 For the analysis, the patients were divided into groups with stage IA and stage IB. Complete remission or complete response was defined as having ≥95% clearance of skin lesions, partial response as 50% to 95% clearance, and nonresponse as less than 50% clearance.2,7 Relapse was defined as a clinically significant disease requiring further therapy. Relapse-free interval or time to relapse was defined as the period between a complete response and relapse.9 Maintenance treatment was any treatment after complete remission.
Data collection from medical records was divided into three parts. Part one was demographic data and history of the disease, including age at diagnosis, gender, underlying diseases and the disease stage. Part two was the treatment protocols from the initial diagnosis to complete remission, including the type of phototherapy (narrowband ultraviolet B or psoralen and ultraviolet A), cumulative dose, number of cumulative sessions, side effects and concomitant therapy. The last part was the maintenance protocols, from the time of complete remission to disease relapse. Treatment regimens were based on patient skin type and disease stage.
The analysis was completed by Cox regression and presented as hazard ratios. The Kaplan–Meier survival curve was performed to plot the relationship between the relapse-free interval and duration of phototherapy during the maintenance period. The cumulative dose and cumulative sessions were also divided into two groups, based on their median values and analyzed as categorical data. Univariate analysis for disease remission and relapse was done for the following variables: age at diagnosis, gender, stage of the disease, cumulative dose and cumulative sessions. Multivariate analysis was done for statistical and clinical reasons. P ≤ 0.05 was considered significant and missing values were not replaced.
Eighty-five patients were diagnosed with an early-stage mycosis fungoides and received phototherapy as the first-line treatment. Twenty-three of them had an incomplete medical record and one of them stopped phototherapy because of side effects before the response was defined. Therefore, 61 patients met the eligibility criteria. Fifty-three patients received narrowband ultraviolet B and eight patients received psoralen and ultraviolet A as an initial treatment.
From this cohort, 57 (93.4%) patients achieved complete remission. Three patients reached partial remission in 6.14 months, 6.34 months and 17.38 months, respectively. One case had no response after receiving the treatment for 8.16 months. The mean age at diagnosis was 44.0 ± 18.8 years, 38 (62.3%) of whom were women. Among these patients, 50 (87.7%) patients were treated with narrowband ultraviolet B and 7 (12.3%) patients with psoralen and ultraviolet A. The disease stage was recorded to be stage IA in 36 patients (67.9%) and IB in 17 patients (32.1%). Median cumulative dose with interquartile range and median cumulative sessions during treatment were 42.9 (22.6–67.7) KJ/cm2 and 38 (30– 67) sessions, respectively. The demographic data and disease characteristics are displayed in Table 1.
|Number of patients||61|
|Age of diagnosis (years; mean±SD)||44.0±18.8|
|Sex, n (%)|
|Disease stage (by clinical), n (%)|
|Disease stage (by pathological), n (%)|
|Patch stage||23 (74.2)|
|Plaque stage||8 (25.8)|
|Patch, n (%)|
|Plaque, n (%)|
|Hypopigmented, n (%)|
|Poikiloderma, n (%)|
|Type of phototherapy, n (%)|
|Narrowband ultraviolet B||53 (86.9)|
|psoralen and ultraviolet A||8 (13.1)|
The total follow-up time to remission was 734.62 person-months. Overall, the median time to complete a response with a 95% confidence interval (95% CI) was 7.8 months (95% CI 6.68–9.54). There was no significant difference between the narrowband ultraviolet B group (7.57 months [95% CI 5.8–8.3]) and the psoralen and ultraviolet A group (9.18 months [95% CI 3.1–28.13]). The overall incidence rate of complete remission was 0.077 (95% CI 0.059–0.100), and the rate was higher in the group treated with narrowband ultraviolet B than in the group treated with psoralen and ultraviolet A (narrowband ultraviolet B 0.08 [95% CI 0.06– 1.00]) versus (psoralen and ultraviolet A 0.058 [95% CI 0.023–0.12]). The incidences rate of complete remission at 3 months and 6 months were 0.06 (95% CI 0.03–0.16) and 0.34 (95% CI 0.24–0.47), respectively. The univariate analysis for predictors of complete remission revealed that the presence of poikiloderma was a predictive factor for a poor response (hazard ratio 0.33, 95% CI 0.12–0.87, P = 0.026), while age, gender and type of phototherapy did not affect the rate of complete remission [Table 2].
|Patient characteristics||Univariate analysis||Multivariate analysis†|
|Hazard ratio||95% CI||P-value||Hazard ratio||95% CI||P-value|
|Age at diagnosis (years)||0.99||0.97–1.00||0.29|
|Type of phototherapy|
|Narrowband ultraviolet B||1||1|
|Psoralen and ultraviolet A||0.66||0.29–1.45||0.31||0.40||0.13–1.29||0.13|
As psoralen and ultraviolet A was mainly administered to patients with more infiltrative disease, such as in cases of plaques or a higher disease stage, adjustments for the following factors were made to eliminate the confounding factors: types of phototherapy, disease stage and initial clinical presentation.10 After the adjustments, poikiloderma and high stage of disease (IB) were found to be associated with a longer time of treatment to remission (20.9 ± 4.35 months in cases with poikiloderma vs 7.57 ± 0.43 months in cases without poikiloderma, P = 0.02 and 7.44 ± 0.46 months for stage IA vs 9.54 ± 2.99 months for stage IB, P = 0.035).
Total follow-up time to relapse was 999.51 person-months. Twenty-nine (50.8%) patients had recurrent mycosis fungoides. Among them, 26 were treated with narrowband ultraviolet B, while three were treated with psoralen and ultraviolet A. Median time of relapse-free interval was 24.78 months (95% CI 15.2–39.6) overall, (24.78 ± 4.67 for narrowband ultraviolet B and 19.9 months for psoralen and ultraviolet A). The overall incidence of relapse was 0.029 (0.034 for narrowband ultraviolet B and 0.012 for psoralen and ultraviolet A).
The univariate analysis had demonstrated that there were no statistically significant differences in age, sex, disease stage, type of phototherapy, cumulative dose and cumulative sessions during the treatment phase between patients with and without relapse.
We further evaluated the effect of maintenance therapy on disease recurrence. Of all our patients, 42 received phototherapy during their maintenance period, while 14 did not (one patient was lost to follow-up after achieving complete response). In the group with phototherapy during the maintenance period (n = 42), more than half still experienced recurrence (27/42 cases). Further details are displayed in Table 3. The analysis found that the longer duration of phototherapy, higher cumulative dose (≥10 KJ/cm2) and higher cumulative sessions (≥25 sessions) throughout the maintenance period could diminish disease recurrence. However, after adjustment for the variable stated above, only maintenance duration showed a significant correlation with a duration of six months or longer significantly increased relapse-free interval period (hazard ratio 0.13, 95% CI 0.04– 0.41, P = 0.001) [Figure 1].
|Maintenance treatment (n=42)||Univariate analysis||Multivariate analysis†|
|Hazard ratio||95% CI||P-value||Hazard ratio||95% CI||P-value|
|Longer duration of phototherapy during maintenance period||0.87||0.82–0.93||<0.001*||0.87||0.82–0.93||<0.001*|
|Cumulative dose (KJ/cm2)|
Of note, topical corticosteroids were widely prescribed during the maintenance phase, 33 cases used topical steroids during maintenance therapy, but there was no significant correlation between their use and disease recurrence rate (P = 0.15). Nevertheless, in the steroid group, a longer median time to recurrence was observed (29.24 ± 3.22 months vs 20.06 ± 0.19 months in nonsteroid group).
Our study has demonstrated a high rate of complete remission (93.4%) in patients receiving either narrowband ultraviolet B or psoralen and ultraviolet A. The results from Table 2 showed different hazard ratios in the multivariable analysis than the univariate analysis in several factors, especially hypopigmented type and patch type. We believe that these hazard ratios from the univariate analysis were confounded by other factors. Therefore, the results from the multivariate model should be more reliable because they were already controlled for the confounding factors.
As found in previous studies, age and gender did not correlate with treatment response.11,12 Several studies in the past have compared the efficacy of treatment between narrowband ultraviolet B and psoralen and ultraviolet A.Diederen et al.,13 Ponte et al.14 and El-Mofty et al.15 reported a higher complete remission rate for patients who received narrowband ultraviolet B compared with that for those who received psoralen and ultraviolet A (81% vs 71%, 68.4% vs 62.1% and 90% vs 80%, respectively), but these were not statistically significant. In contrast, Ahmad et al.9 and Almohideb et al.11 reported that psoralen and ultraviolet A had a higher complete remission rate than narrowband ultraviolet B. However, their results did not show any statistical significance either. Recently, a study of 227 cases with early mycosis fungoides10 concluded that psoralen and ultraviolet A led to a significantly higher complete response rate (77.1% in psoralen and ultraviolet A vs 53.8% in narrowband ultraviolet B, P = 0.001). The superiority of psoralen and ultraviolet A was later emphasized in a meta-analysis of seven studies, comparing the efficacy and adverse effects of psoralen and ultraviolet A and narrowband ultraviolet B.6 Altogether, 778 cases with early-stage mycosis fungoides (stage IA - IIB) were studied with at least 10 patients in each comparator group. The analysis found a significantly higher complete response in psoralen and ultraviolet A (73.8% vs 62.2%, P = 0.04) across all disease stages. Our study suggests otherwise with no significant difference in the rates of complete remission between the two groups, even after adjustments for disease stage and clinical presentation. This could be explained by the small number of subjects in the psoralen and ultraviolet A group (n = 7).
In terms of adverse reactions, mild erythema following phototherapy was reported in only three cases, one from psoralen and ultraviolet A and two from narrowband ultraviolet B. In the meta-analysis, no significant difference in adverse events were found between psoralen and ultraviolet A and narrowband ultraviolet B, but the percentages of erythema, nausea, phototoxic effects and dyspepsia were frequently higher in psoralen and ultraviolet A.6,16,17 In addition, nonmelanoma skin cancers were reported in some psoriatic patients with psoralen and ultraviolet A therapy in a dose-dependent manner, occurring when the cumulative dose exceeded 2000 J/cm2 or 200 sessions.18,19 With narrowband ultraviolet B, patients do not have to take psoralen or use sun-protective methods.6 There have not been any reports on the association with skin cancers.
Patients’ initial clinical presentation may also affect the treatment response. Our study reveals that poikiloderma is a predictive factor for poor response. In our study, poikiloderma was found in seven patients, of which five (71%) achieved complete remission. Although the rate of complete response was high, the median time to remission in patients who had poikiloderma was significantly longer (20.9 ± 4.35 months vs 7.57 ± 0.43 months in those without poikiloderma). The results might guide physicians to choose a more aggressive treatment for patients with poikiloderma, such as psoralen and ultraviolet A or combination therapy. Abbott et al.20 conducted a retrospective study to improve the data availability of poikiloderma in mycosis fungoides and concluded that 24/27 (89%) patients with poikiloderma responded to phototherapy, but they did not mention the time to remission. Another clinical feature known to affect treatment response is rash morphology. Gökdemir et al. found that the patch presentation has a significantly higher rate of complete remission when compared with the plaque type (P = 0.04).21 However, the difference was not highlighted in our study.
Disease recurrence in early-stage mycosis fungoides after narrowband ultraviolet B or psoralen and ultraviolet A treatment still poses a challenge. The percentage of relapse in our study was up to 50.8%, but the number of studies on the predictive factors for relapse is small. Felix et al. conducted a study to evaluate the long-term disease-free survival after narrowband ultraviolet B treatment in stage I mycosis fungoides and its predictive factors.22 The study found that younger age (<50 years) and earlier stage (stage IA) had a significant effect on longer disease-free survival time. In contrast with other studies,11,23 we failed to establish the correlation between disease stage and recurrence. Age, sex and cumulative sessions during treatment also did not associate with disease recurrence in our study which agreed with another study.11 When compared with narrowband ultraviolet B, psoralen and ultraviolet A had a lower incidence of relapse but a shorter time to relapse, also inconsistent with the previous data.6
Currently, a guideline for the dose and duration of maintenance phototherapy is still based on expert opinion. Previous studies have compared the relapse-free interval directly between patients who received and did not receive maintenance therapy.10,24-26 All of them reported no significant difference in relapse-free interval between the two groups. For the patients who did not receive maintenance phototherapy, the relapse ranged from 29% to 100% with a mean relapse-free interval between 5.9 and 14.5 months.2 Nikolaou et al. reported that maintenance therapy does not associate with a reduction in relapse rate (P = 0.51) but do prolong the time to relapse (32.3 months in maintenance vs 19.5 months in nonmaintenance, P < 0.002).10 Similarly, the maintenance treatment in our study had no role in reduction of relapse rate, even after adjustment for the disease stage, but a duration of six months or longer could increase the relapse-free interval period. It must be noted that most (73.7%) of the patients in our center received phototherapy during their maintenance period.
The limitation of our study is that it is a retrospective study with a small number of subjects. In the future, stronger evidence-based research, such as randomized control trial or meta-analysis, may help clarify the subject.
Narrowband ultraviolet B and psoralen and ultraviolet A are suitable options for the treatment of early-stage mycosis fungoides. The presence of poikiloderma may indicate a predictive factor for poor prognosis and a longer time to complete remission. The relapse rate in our study was high, occurring around two years after complete remission. Maintenance phototherapy after disease remission for at least six months can significantly prolong the relapse-free interval period. In addition, topical steroid usage during this period might help prolong the time to relapse.
This study was supported by the Division of Dermatology, Faculty of Medicine, Ramathibodi Hospital, Mahidol University, Bangkok, Thailand.
Declaration of patient consent
The authors certify that they have obtained all appropriate patient consent.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.