Kaplan-Meier actuarial survival curves of patients with clinical stage IA and IB/IIA mycosis fungoides treated with psoralen UV-A therapy (P = .10).
Kaplan-Meier disease-free survival curves of patients with stage IA and stage IB/IIA mycosis fungoides who achieved complete remission with psoralen plus UV-A therapy (P = .08).
Kaplan-Meier overall survival curves of relapse and nonrelapse patients (P = .93).
Querfeld C, Rosen ST, Kuzel TM, Kirby KA, Roenigk HH, Prinz BM, Guitart J. Long-term Follow-up of Patients With Early-Stage Cutaneous T-Cell Lymphoma Who Achieved Complete Remission With Psoralen Plus UV-A Monotherapy. Arch Dermatol. 2005;141(3):305–311. doi:10.1001/archpedi.161.4.356
To evaluate long-term outcomes, impact of maintenance therapy and potential curability of patients with mycosis fungoides (MF) treated with psoralen plus UV-A (PUVA) monotherapy.
Single-center retrospective cohort analysis.
Academic referral center for cutaneous lymphoma.
A total of 66 of 104 patients with clinical stages IA to IIA MF who achieved complete remission (CR) after PUVA monotherapy between 1979 and 1995.
Main Outcome Measures
Kaplan-Meier actuarial survival and disease-free survival curves were compared between stage IA and IB/IIA cases. Patients were stratified into relapse and nonrelapse groups based on whether their MF relapsed during study follow-up. Baseline characteristics and treatment were compared between these groups.
Median follow-up time was 94 months (range, 5-242 months). Thirty-three patients maintained CR for 84 months (range, 5-238 months), and 33 patients experienced relapse with a median disease-free interval of 39 months (range, 2-127 months). There was no significant difference in baseline characteristics between patients in the nonrelapse and relapse groups. Those in the nonrelapse group received a higher cumulative dosage to CR (P = .03) and required longer treatment periods to achieve CR (P = .03). Disease-free survival rates at 5 and 10 years for all patients with stage IA were 56% and 30%, respectively, and for stage IB/IIA, 74% and 50%. Actuarial survival rates at 5, 10, and 15 years were 94%, 82%, and 82%, respectively, in patients with stage IA, and 80%, 69%, and 58% in patients with stage IB/IIA. The overall survival rate for the nonrelapse and relapse groups did not show any statistical difference. One third of the patients developed signs of chronic photodamage and secondary cutaneous malignancies.
Psoralen UV-A is an effective treatment for MF, inducing long-term remissions and perhaps in some cases disease “cure.” Thirty percent to 50% of patients remain disease free for 10 years, but late relapses occur. Long-term survival is not affected by relapse status, and the risk of photodamage needs to be measured against the possible benefit of greater disease elimination.
Cutaneous T-cell lymphomas (CTCLs) represent a spectrum of non-Hodgkin lymphoproliferative disorders characterized by skin-homing lymphocytes.1 The most common type is the epidermotropic variant, mycosis fungoides (MF),2 which is usually characterized by chronic, indolent progression, but a limited number of patients develop biologically aggressive disease with a poor prognosis.3 At present, no curative therapy exists. In early-stage CTCL, the disease is mostly limited to cutaneous infiltrates; therefore, most of these cases can be treated successfully with topical therapies.4,5
Psoralen plus UV-A exposure (PUVA) is beneficial for the treatment of various dermatoses.6 In 1976, PUVA was introduced for the treatment of MF and shortly thereafter became one of the preferred treatment options.7 Several studies have confirmed high remission rates in early stages of MF, including complete remissions (CRs) in up to 71.4% of patients.8- 14 In combination with interferon alfa, PUVA therapy produces significant benefit in stage IIB or more advanced disease.15- 17
Besides short-term adverse effects, including nausea from ingestion of psoralen, erythema, pruritus, and photodermatitis, long-term high-dose PUVA therapy is associated with an increased risk of skin malignancies such as squamous or basal cell carcinoma.18,19 The risk of malignant melanoma is controversial. It was found to be increased in a single cohort of PUVA-treated patients about 15 years after their first PUVA treatment and with higher doses of PUVA, but this risk could not be confirmed by other studies.20,21
However, practical considerations and the significant efficacy of this treatment as confirmed by investigative studies make PUVA an attractive choice. Narrowband UV-B phototherapy is considered to be less carcinogenic and may be an alternative treatment option in early-stage MF.22- 24 However, the remission time is short, and a maintenance schedule is difficult to establish. A report with a limited number of patients suggests that UV-A1 phototherapy may also be effective.25 Other attempts to reduce PUVA exposure and minimize adverse effects have led to the use of combined regimens with retinoids and interferon alfa. No randomized trials have confirmed similar efficacy or improved toxicity profiles.
Despite the documented short-term benefits of PUVA therapy, few data exist on the long-term course of PUVA-treated patients and potential curability.9,11 In the present retrospective study, we evaluate long-term outcomes, impact of maintenance therapy, and possible disease curability for patients with early-stage IA to IIA CTCL who achieved CR with PUVA therapy alone.
Between 1979 and 1995, a total of 104 patients diagnosed as having MF stage IA to IIA and treated with PUVA monotherapy were retrospectively identified from the records at the Multidisciplinary Cutaneous Lymphoma Clinic at the Feinberg School of Medicine of Northwestern University. Fifty-seven patients with stage IA, 42 patients with stage IB, and 5 patients with stage IIA were identified. For classification and staging, patients underwent complete physical examination, skin biopsies for routine histopathologic evaluation, complete blood cell count with differential and Sézary cell evaluation, comprehensive chemistry analysis, and lymph node biopsy in clinically abnormal lymph nodes. In selected instances, additional studies included chest radiography or computed tomography. All patients were staged at presentation at our institution according to the TNMB classification.26 Diagnosis of MF was confirmed by histologic analysis, which was repeated after disease remission or relapse.
From our database for cutaneous lymphomas, we identified 66 patients with clinical stage IA to IIA disease who achieved CR after treatment with PUVA. Thirty-nine patients had stage IA disease, 22 had stage IB disease, and 5 had stage IIA disease. Forty-nine patients were observed for a minimum of 60 months. The remaining 17 patients were either lost to follow-up (n = 10) or died (n = 7) within this period, having a median follow-up time of 29 months (range, 5-51 months). Patient data were obtained after the study met the approval by the local institutional review boards.
Patients ingested methoxsalen 2 hours prior to UV-A exposure. Initial exposure times were limited to skin phototype according to Fitzpatrick grading system, the ability to tan, and the patient’s history of sunburns. The initial UV-A dose was approximately 0.5 J/cm2 and was increased by 0.5 J/cm2 per treatment or up to the minimal erythema dose. Therapy was typically given 2 to 3 times a week until CR was achieved. Additional maintenance therapy was performed in most patients after complete clearing of cutaneous lesions. The number of treatments was gradually reduced to between one per week and one every 6 weeks at a constant dose, as described elsewhere.27 For patients who experienced relapse during maintenance therapy, the frequency of treatment was increased to 1 to 3 times weekly. Patients were allowed to use emollients, and 3 patients used mild-potency topical steroids in areas inaccessible to UV-A radiation. A baseline ophthalmologic examination was performed before initiation of treatment.
Complete remission was defined as no evidence of disease for a minimum of 4 weeks with required histologic confirmation. Progression of disease was defined as recurrence of disease and worsening to higher-stage disease. The duration of disease-free survival was determined as the interval between initial documentation of CR and an event (eg, a patient’s disease recurred, a patient’s course of treatment changed, complications occurred, or a patient died). Relapse was defined as the return of clinically significant disease requiring resumption of more intensive PUVA or alternative therapy.
All statistical analyses were based on 66 eligible patients. Patients were stratified into relapse and nonrelapse groups, based on whether they experienced disease relapse during study follow-up. Baseline variables for each group were compared using Fisher exact, χ2, and Wilcoxon rank-sum tests. Follow-up time and disease-free survival time were calculated from the date of CR, and differences between the groups were compared with the Wilcoxon rank-sum test. Kaplan-Meier estimates of overall survival and disease-free survival were calculated. Survival functions among patients with stage IA vs IB/IIA were compared using the log rank test. Significance was defined as P<.05.
The clinical characteristics of 66 patients who were treated with PUVA are summarized in Table 1. Thirty-three patients had no relapse and 33 patients experienced relapse after CR. There were no statistically significant differences between these groups. Median follow-up time of the nonrelapse group was 84 months vs 99 months for the relapse group. Patient age at presentation ranged from 22 to 89 years; median age was 57 years for the nonrelapse group and 54 years for the relapse group; and female-male ratios were 1:1.4 and 1:1.8, respectively. Ninety-four percent of our patients were white. Most patients had clinical stage IA disease at presentation. Of the 33 nonrelapse patients, 16 (48%) had stage IA, 14 (42%) had stage IB, and 3 (9%) had stage IIA disease. Of the 33 relapse patients, 23 (70%) had stage IA, 8 (24%) had stage IB, and 2 (6%) had stage IIA disease (P = .11).
Previous treatments included topical steroids in 45 patients, topical nitrogen mustard in 3 patients, PUVA or UV-B in 6 patients, and treatment with local electron beam, systemic steroid, interferon alfa, and CHOP (cyclophosphamide, doxorubicin hydrochloride, vincristine sulfate, and prednisone) in 1 patient each. The median number of prior treatments for each group was 1 (data not shown). Twelve patients (36%) in the nonrelapse group and 3 (9%) in the relapse group had no previous therapy for CTCL. Three (9%) of the nonrelapse patients received 1 prior systemic treatment vs no systemic treatment in the relapse group.
The results of PUVA and maintenance therapy are summarized in Table 2. The median time to CR was 4 months for the nonrelapse patient group vs 2 months for the relapse patient group (P = .03). Nonrelapse patients received a higher cumulative dose to CR (P = .03). This was not attributed to the different ratio of stage IA to IB/IIA disease compared with the relapse group. The cumulative dose by stage was also higher for nonrelapse patients compared with relapse patients. Patients with stage IA received a higher cumulative dose than patients with stage IB/IIA. Clinical stage did not reveal any significant effect on relapse status (P = .07). There were no statistical differences in highest single dose per treatment.
The effect of maintenance treatment with PUVA was evaluated in the 2 patient cohorts. Thirty patients (91%) in the relapse group and 31 (94%) in the nonrelapse group received maintenance therapy. Most of the patients were gradually tapered from once every week to once every 3 weeks. The comparative results revealed that the number of patients in higher maintenance intervals were too small to perform any significant statistical analysis. Five patients did not receive a maintenance regimen; 3 of these experienced relapse at 5, 61, and 127 months, respectively. The 2 nonrelapse patients maintained CR for 24 and 86 months, respectively.
The overall survival rates of the 66 patients grouped with clinical stage IA and IB/IIA are shown in Figure 1. The median survival of patients with stage IA and IB/IIA has not been reached; however, the actuarial survival rates at 5, 10, and 15 years were 94%, 82%, and 82%, respectively, in patients with stage IA disease and 80%, 69%, and 58%, respectively, in patients with stage IB/IIA disease. The median follow-up time for all patients was 94 months (range, 5-242 months). Disease-free survival rates at 5 and 10 years for stage IA disease were 56% and 30%, respectively; for stage IB/IIA, 74% and 50% (Figure 2). The disease-free survival curves show that patients with stage IB/IIA disease showed a trend toward more favorable disease-free survival than patients with stage IA disease, but the difference did not quite reach statistical significance (P = .08). There was no statistical difference in overall survival between the nonrelapse and relapse cohorts (Figure 3, P = .93).
The median time of disease-free survival for the 33 nonrelapse patients was 84 months compared with 39 months for the 33 relapse patients (Table 3). Seven (21%) of the relapses occurred within the first 2 years of CR, and 26 (79%) occurred more than 2 years after CR. In 3 patients with stage IA (9%), disease progressed to higher T stage (T2), and in 4 patients (12%) with stage IB/IIA, it progressed into higher T stage (data not shown). Twelve patents in the relapse group (36%) remained disease-free at last follow-up after resumption of more intensive PUVA therapy. A few patients experienced multiple relapses but continued to respond to repeated regimens of PUVA treatment. Eight (24%) of the 33 relapse patients died, but only 4 deaths were attributable to MF. Other causes were metastatic melanoma, colon cancer, or unknown. Six patients (18%) died in the nonrelapse cohort. Causes were lung cancer, soft tissue sarcoma, or unknown.
In general, PUVA treatment was well tolerated. Most patients developed mild short-term adverse effects such as erythema, pruritus, and nausea (Table 4). Two patients developed polymorphous light eruption. Corneal abrasion and gastrointestinal bleeding occurred in 1 patient each. All patients were able to continue therapy with decreased PUVA dose or treatment interruptions. Signs of chronic photodamage including telangiectasia, xerosis, solar elastosis, solar lentigines, and lentiginous nevi were found in 18 patients (27%).
Seventeen patients (26%) developed skin cancer that was histologically identified as Bowen disease or squamous or basal cell carcinoma. One patient each developed melanoma in situ and metastatic melanoma during maintenance treatment. Of the 12 patients who developed nonmelanoma skin cancer, 9 had received prior treatment with topical steroids and 1 each with PUVA and interferon alfa. One patient received combined topical treatment with steroids and nitrogen mustard.
The skin-targeted treatments for early-stage IA to IIA CTCL include PUVA, narrowband UV-B, total skin electron-beam irradiation, as well as topical preparations of steroids, retinoids, carmustine, or nitrogen mustard. It has been hypothesized that in early-stage CTCL, clinically detectable disease is mostly confined to the skin, and therefore, most of these cases can be treated successfully with topical applications with low risk of systemic complications. Results of PUVA therapy have been reported, but only a few studies (with limited numbers of patients with CTCL and prelymphomatous conditions such as parapsoriasis en plaque) provide long-term follow-up information.
Northwestern University had one of the first PUVA units in the United States, and our Multidisciplinary Cutaneous Lymphoma Group has been a leader in the use of PUVA for treatment of CTCL to achieve remission from disease, offer palliation of symptoms, and prevent disease progression. From 1979 to 1995, we have treated 104 patients with stage IA to IIA MF with PUVA monotherapy. Although our group has shown a high remission rate in patients with early MF (63%), the long-term follow-up of these patients who achieved CR is not known.10 In the present study, we present the follow-up data of 66 patients who achieved CR between 1979 and 1995.
The results of our retrospective study show that 33 patients (50%) sustained remissions. Fifty percent experienced relapse, but interestingly, after relapse most patients again responded when PUVA treatment was resumed more intensively, and 12 relapse patients (36%) remained disease free at last follow-up. A few patients experienced multiple relapses but continued to respond to repeated regimens of PUVA.
The comparison of our results with previously reported PUVA studies is problematic because of differences in diagnostic and response criteria for CTCL, patient selection, PUVA dosing, and the use of a maintenance regimen. Two reports from our center, which included patients with parapsoriasis and advanced stages of disease, showed mean response durations of 29.5 and 43 months, respectively, during respective mean and median follow-up periods of 38 and 43 months.10,28 Relapse patients had a mean response duration of approximately 6.3 months.28 Hönigsmann et al11 reported a mean disease-free interval of 20 months for stage IA disease and 17 months for stage IB disease during a mean follow-up of 44 months.11 The Scandinavian Mycosis Fungoides Study Group12 and other, mostly small, studies showed similar results.13,14 Shorter disease-free durations reported in these studies might result from inclusion of patients with later-stage disease. They might also be owing to differences in PUVA dose adjustments, clinical appearance of lesions, number of patients, and follow-up time. The present study focuses on patients with early-stage CTCL and has a significantly longer follow-up time (median, 94 months).
We found that nonrelapse patients required a significantly higher cumulative dose to achieve CR than relapse patients (P = .03) and required longer treatment periods to attain CR (P = .03). Clinical stage had no significant impact on relapse status. The more intense therapy may have enhanced the remission from induction treatment. Alternatively, there may be a biological difference in the neoplastic lymphocytes of patients who respond slowly but completely compared with patients who respond rapidly but not durably. Patients with stage IA disease in both cohorts received a higher cumulative dose than those with stage IB/IIA disease. The reason patients with limited skin involvement require higher doses is not known, but this corresponds to our clinical observation that patients with limited skin involvement appear to be more resistant to topical treatments.The benefit of maintenance treatment remains controversial without randomized clinical trial data. While almost all patients received maintenance treatment, 50% experiences relapse while 50% showed a long-lasting response. The effect of maintenance treatment was evaluated in nonrelapse and relapse patients, but the number of patients in each cohort was too small to evaluate the impact of different intervals or duration of maintenance.
After treatment, skin samples were considered in histologic remission when atypical lymphocytes were absent. However, a minimal submicroscopic residual disease cannot be excluded. Thus, patients might have had their maintenance regimen discontinued prematurely. In addition, UV-A (320-400 nm), with its peak emission wavelength between 330 and 340 nm, penetrates the skin approximately 1 to 2 cm into the mid dermis. Deeper dermal structures such as follicles involved by malignant lymphocytes might not benefit from this treatment. T-cell receptor gene rearrangement analysis by polymerase chain reaction has been recently suggested to detect minimal residual disease, but its prognostic value has not been determined yet.29
Given data that suggest that circulating neoplastic lymphocytes are present even in early-stage CTCL, long-lasting remissions presumably require both topical and systemic effects.30,31 Various mechanisms have been proposed to explain the mode of therapeutic action of PUVA. In vitro studies have shown that peripheral blood mononuclear cells undergo apoptosis after exposure to PUVA, and immunostimulating effects after exposure may be additional mechanisms for crucial benefit in early-stage disease.32,33 It is known that in early-stage MF, lymphocytes require signals from Langerhans cells and other antigen-presenting cells. Psoralen plus UV-A exposure reduces the number of epidermal antigen-presenting Langerhans cells, which suggests that depletion of Langerhans cell population might lead to additional growth retardation of CTCL.34,35
An unexpected finding was the trend for long-term disease-free survival rates to be higher in patients with stage IB/IIA disease than in those with stage IA. However, actuarial survival curves revealed that while response duration was better maintained in patients with stage IB/IIA disease, overall survival was better in patients with stage IA, which is consistent with prior longitudinal studies of the natural history of MF.36 Our long-term outcome results compare favorably with recent published experiences with other skin-directed therapies, such as topical nitrogen mustard chemotherapy or total skin electron-beam irradiation. Reported overall survival rates for nitrogen mustard were 97% and 69% at 5 and 15 years, respectively, for patients with T1 disease, and 72% and 48%, respectively, for patients with T2 disease and are similar to our actuarial survival rates.37 Total skin electron-beam irradiation showed 88% overall survival for patients with stage IA and 76% for patients with stage IB disease at 15 years.38 However, the disease-free survival rates of both treatments in Jones et al38 were more favorable in patients with stage IA disease than they are in the present study.
Long-term remissions of more then 8 years, suggesting a possible “cure,” have been reported in a limited number of patients treated with topical nitrogen mustard chemotherapy.39 We were interested to see how these data compared to our results with PUVA. Thirty percent of our patients with clinical stage IA and 50% of those with clinical stage IB/IIA remain in continuous remission approximately 10 years after initial treatment. Despite the long-term remissions in some patients, all relapses for patients with stage IB/IIA have been documented in less than 10 years. However, for patients with stage IA disease, relapses occurred throughout the observation period without evidence of a plateau in the tail of the curve (Figure 2). This finding indicates that PUVA may induce long-lasting disease-free intervals for early stages, but it does not indicate whether permanent disease eradication can be achieved.
The most common short-term adverse effects were erythema, pruritus, and nausea, which were usually mild at presentation and generally manageable with dose adjustments of UV-A or psoralen or treatment interruptions. Psoralen plus UV-A treatment is known to have potent mutagenic and carcinogenic effects. Approximately one third of our patients developed chronic photodamage changes, and one third developed skin cancer. Nine of 15 patients with nonmelanoma skin cancer had 1 prior treatment with potent topical steroids. Only 2 patients had prior potentially skin-damaging therapies.
Psoralen plus UV-A treatment is an effective monotherapy for early-stage CTCL with potential for long-term remissions if the patient continues with maintenance treatment. Patients with disease recurrence often responded rapidly to repeated courses of therapy. Long-term survival is not affected by relapse status. Studies probing the concept of cure in CTCL are still controversial, since even initial aggressive chemotherapy has not been shown to improve prognosis.40- 42 Our data indicate that 30% to 50% of patients with early-stage MF remain disease free for close to 10 years, but based on the occurrence of late relapses, a permanent cure is not suggested. Our results demonstrate that PUVA dose and duration might have an impact on disease-free survival, and the risk of photodamage needs to be weighed against the possible benefit of greater disease suppression.
Correspondence: Joan Guitart, MD, Department of Dermatology, Northwestern University Medical School, 222 E Superior St, Suite 220, Chicago, IL 60611 (firstname.lastname@example.org).
Accepted for Publication: June 4, 2004.
Previous Presentation: This study was presented at the 65th Annual Meeting of the Society of Investigative Dermatology and at the International Society for Cutaneous T-Cell Lymphoma; April 28-May 1, 2004; Providence, RI.
Financial Disclosure: None.