Introduction

Ocular adnexal lymphoma (OAL) occurs in approximately 1–2 % of non-Hodgkin lymphomas (NHL) and 8 % of extranodal lymphomas [1]. The incidence is increasing by approximately 6.3 % annually, according to Surveillance, Epidemiology, and End Results data [2]. In Korea, the frequency of OAL is higher than in Western countries, and younger patients are affected [1, 3, 4]. The major histology is marginal zone B cell lymphoma of the mucosa-associated lymphoid tissue type (MALT), followed by follicular, diffuse large B cell, mantle cell, and small lymphocyticlymphoma in rare cases. Typically, marginal zone B cell lymphoma involving ocular adnexa is a clinically slow growing tumor in the eyelid, orbit, lacrimal gland, or conjunctiva, and the slow growth may contribute to delayed diagnosis.

Chronic antigen stimulation is considered to be a pathogenic mechanism causing malignant transformation of a normal lymphoid cell to MALT lymphoma. Chlamydophila psittaci (Cp) has been suggested as an antigen in OAL [5]. In humans, Cp is an etiological agent of psittacosis, which is caused by exposure to infected animals including birds, domestic mammals, and pets. The risk of OAL is markedly increased by contact with animals, and Cp DNA has been found in 80 % of OAL patients in Italian studies [5, 6]. However, the rate of Cp positivity varies between geographical regions and also between studies from the same country [7, 8]. A report from Korea showed that the Cp DNA detection rate from paraffin-embedded tissues was 79 % in 33 OAL cases, similar to reports from Italy [9, 10].

For patients with localized OAL, radiotherapy confers a high control rate with long-term efficacy. However, radiation to the involved eye can result in complications, including cataract (30–50 %), xerophthalmia (20–40 %), and retinopathy (2 %) [1, 11, 12]. Other options, such as single-agent chemotherapy, immunotherapy with monoclonal anti-CD20 antibody, and antibiotic therapy, have been studied in OAL [1, 13]. Pilot and extended multicenter trials of doxycycline treatment to eradicate Cp have been reported by Ferreri et al. [14, 15]. These authors recently published efficacy and safety data of doxycycline treatment of 47 newly diagnosed OAL patients in an international phase II trial. After a median follow-up of 37 months, the 5-year progression-free survival (PFS) was 55 ± 9 % [16].

We previously reported the efficacy of first-line therapy with doxycycline in 38 patients with newly diagnosed OAL at our institution. After a median follow-up of 26.4 months, doxycycline treatment resulted in a 3-year time-to-treatment failure rate of 84 % [9]. In the current study, we extended the follow-up duration in a larger number of patients who were newly diagnosed with OAL of the MALT type and treated up-front with doxycycline.

Patients and methods

Patients

Doxycycline was given as a first-line treatment to 90 patients who had been newly diagnosed with ocular adnexal marginal zone B cell lymphoma at the Seoul National University Hospital between January 2004 and December 2012. Any patient who received combination chemotherapy or radiotherapy in advance was excluded. All study patients were classified according to the American Joint Committee on Cancer (AJCC) tumor, nodes, and metastases (TNM)-based staging system, with the T classification categorized as follows: T1, conjunctival involvement; T2, involvement of the orbit; T3, infiltration of preseptal eyelid tissues; and T4, extension beyond orbit to adjacent structures, such as bone and brain (Fig. 1). This study was approved by the Institutional Review Board at Seoul National University Hospital (H-1403-003-560).

Fig. 1
figure 1

Computed tomography findings of patients with ocular adnexal lymphoma according to T (tumor) classification

Full size image

Treatments and response evaluation

Each patient was treated with one or two cycles of doxycycline (100 mg bid) for 3 weeks. Doxycycline outcomes were evaluated in patients who were followed for more than 6 months after completion of therapy. Medical history, ophthalmologic examination with anterior segment photography by experienced ophthalmologists, orbital magnetic resonance imaging (MRI) or computerized tomography (CT) scan of the lesion, physical examinations, complete blood count, and chemical tests including lactate dehydrogenase (LDH) were performed and recorded at baseline. An ophthalmologic examination, complete blood count, chemistry including LDH, and radiologic examination (CT or MRI) were performed every 6 months. Patients who progressed after doxycycline treatment received second-line treatment. Second-line treatment was cyclophosphamide, vincristine, and prednisolone (CVP) combination chemotherapy and/or radiotherapy (30.6 Gy), as decided by the clinician and patient. Thirty-one patients who received second-line therapy assessed response using modified international workshop criteria [17]. Response was defined as follows: complete remission (CR), complete disappearance of all detectable ophthalmic and radiographic evidence of disease and eye-related symptoms if present before therapy; partial remission (PR), 50 % or more decrease in the sum of the product of the greatest diameters; progressive disease (PD), any new lesion or 50 % or more increase from the smallest sum of the product of the greatest diameters; stable disease (SD), an absence of CR, PR, and PD.

Statistical analysis

The primary study end point was PFS, defined as the time elapsed between doxycycline initiation and tumor progression or censoring of patients at the date of last follow-up. All of the study patients were alive at the last follow-up. Survival curves for PFS were determined by the Kaplan-Meier method [18] and compared with results of log-rank tests. Univariate and multivariate analyses of PFS according to clinical factors were performed with a Cox proportional hazards regression model [19]. Absolute lymphocyte and neutrophil counts within 1 week before doxycycline treatment were used to define lymphocytosis and neutrophilia. Lymphocytosis or neutrophilia was defined as an absolute lymphocyte or neutrophil count higher than the respective median value. Two-sided P values <0.05 were considered to indicate statistical significance. All statistical analyses were performed using SPSS statistics software (version 20) (IBM Corp.; Chicago, IL, USA).

Results

Patients

Clinical characteristics of the 90 study patients are listed in Table 1. The median age at diagnosis was 48 years (range, 21–71 years) with a male-to-female ratio of 1:1.6, and all patients had Eastern Cooperative Oncology Group (ECOG) performance status 0–1. The most common site of presentation was conjunctiva (82 %), followed by orbit (13 %), eyelid (3 %), and lacrimal gland (2 %). Serum LDH was elevated in 16 % of patients, and only one patient had B symptoms at diagnosis. Two cycles of doxycycline were administered in 80 % of the patients. Measurable lesions were seen on CT or MRI in 42 % of patients. According to the TNM staging, 62 patients (69 %) were T1N0M0, 22 (24 %) were T2N0-2 M0, 5 (6 %) were T3N0-2 M0, and 1 (1 %) was T4N0M0. Median absolute lymphocyte and neutrophil count was 1.95 × 109/L and 3.01 × 109/L, respectively.

Table 1 Patient characteristics
Full size table

Treatment outcomes

All patients were alive at the time of analysis (May 2013). During the median follow-up period of 40.5 months, median PFS was not reached. The 5-year PFS was 60.9 % by intent-to-treat analysis. Twenty-four patients (27 %) responded to doxycycline and 34 patients achieved stable disease (Table 2). Thirty-one patients (34 %) showed local treatment failure without systemic dissemination. In these patients, median time to treatment failure after doxycycline treatment was 4 months (range, 1–56 months). There was no serious adverse event reported during doxycycline therapy. Thirty-one patients received salvage treatment. Twenty patients who received CVP after doxycycline failure showed complete response in six patients, partial response in 11 patients, and SD in three patients, without any PD. Nine patients who progressed after doxycycline and two patients who had SD with intolerable eye symptoms received radiotherapy. Every patient achieved response, with CR in 9 and PR in 2 patients (Table 3). More patients received CVP chemotherapy in patients with >T1 stage than patients with stage T1. The 5-year PFS in patients treated with salvage radiotherapy and/or CVP was 100 %. Examination of PFS according to TNM staging showed that median PFS in T1N0M0 patients was significantly better than in patients at more advanced T stages (median survival not reached vs. 20 months, log-rank P < 0.0001; see Fig. 2).

Table 2 Tumor response to doxycycline by T stage
Full size table
Table 3 Response to second-line therapy
Full size table
Fig. 2
figure 2

Progression-free survival according to the TNM staging

Full size image

Survival analysis and predictive factors

We further analyzed PFS according to clinical variables, such as gender, age, LDH level, laterality, doxycycline cycle, lymphocytosis, neutrophilia, and TNM staging. In univariate analysis, doxycycline cycle and TNM staging were statistically significant (Table 4). There was no significant association between number of cycles and TNM staging (chi-square, P = 0.227). In multivariate analysis, the hazard ratio (HR) of patients with a more advanced stage than T1N0M0 compared to patients with T1N0M0 was 4.35 (95 % confidence interval [CI], 2.03–9.32; P < 0.001), and the HR of two cycles of doxycycline treatment compared to a single cycle was 0.31 (95 % CI, 0.14–0.60; P = 0.004) (Table 4).

Table 4 Univariate and multivariate analysis of factors associated with PFS after doxycycline
Full size table

Discussion

In this long-term large cohort study, almost two-thirds of patients with OAL remained progression-free 5 years after doxycycline treatment. In addition, none of the patients had disease progression with salvage chemotherapy and/or radiotherapy after doxycycline failure. Our results suggest that patients with low-grade localized OAL can be initially treated with doxycycline without serious complications or compromising the salvage treatment after progression or doxycycline failure. This finding is similar to that of a phase-II trial that reported a 5-year PFS rate of approximately 68 % in 14 patients who achieved Cp eradication after doxycycline treatment of 100 mg twice daily for 3 weeks [16]. The PFS rate in this study is comparable to the results from first-line CVP chemotherapy at our institution [20]. Because our previous data showed that patients who received two cycles of treatment had a higher response rate and rapid response [9], we administered two cycles of doxycycline 100 mg twice daily for 3 weeks as standard in this study. However, patients who complained of eye-related symptoms consistently, or those who showed early disease progression after only one cycle of doxycycline, received salvage therapy in advance.

The geographical difference in Cp infection rate is an important factor for interpretation of discrepancies between results. The reported Cp infection rate is higher in Korea (79 %) and Italy (87 %) than in Austria (54 %) [7]. However, Cp DNA was not detected in OAL specimens in Japanese patients [21, 22]. The lack of benefit of one-cycle doxycycline treatment in a blinded study of 11 Austrian patients may have resulted from the low Cp infection and eradiation rates [23]. Recently, Ferreri et al. showed that Cp eradication was associated with better response rate and 5-year PFS [16].

The Ann Arbor staging system cannot differentiate patients who could benefit more from doxycycline treatment. Therefore, we investigated study patients according to the TNM staging system developed by the American Joint Committee on Cancer [24]. Feasibility of TNM staging was previously studied in 50 Korean patients, and the T1N0M0 group showed a higher 3-year PFS than patients with a more advanced stage [25]. In this study, TNM staging was a significant predictive factor for PFS in univariate and multivariate analysis. This suggests that patients who are staged as T1N0M0 might be good candidates for up-front doxycycline treatment.

All of the patients who failed doxycycline received radiotherapy and/or chemotherapy as salvage treatment. The response rate in these patients was 100 % for radiotherapy and 85 % for CVP. The efficacy of salvage radiotherapy was comparable to that of first-line radiotherapy from previous reports [1], although the response rate of salvage CVP chemotherapy seems to be inferior to that of first-line CVP that we previously reported [20]. There can be a potential bias in efficacy of CVP because the patients who had >T1 disease received more CVP than those who had T1 disease.

It is known that younger patients (mean age 45.9 years) are predominantly affected in Korea [3], and the median age in our study (48 years) was very close to this mean value. Older age has been reported to be a poor prognostic factor for disease-free survival and overall survival in patients with OAL [4, 26, 27]. As younger patients have a better prognosis, a treatment modality that is safe and less complicated, such as doxycycline, could be preferable to radiotherapy or chemotherapy as a first-line therapy. A recent review article has discussed clarithromycin for targeting the Cp infection in patients with OAL; however, the data is limited [28].

Although Cp DNA was not examined in this study, the prevalence of Cp infection has been reported to be approximately 60–80 % in Korean patients with OAL, including results from our institution [9, 10]. For future studies, the feasibility of Cp DNA detection in small biopsy samples should be considered [29].

In conclusion, first-line treatment with doxycycline in patients with localized OAL of MALT is effective without causing serious toxicities. Especially, patients of T1N0M0 stage could be considered good candidates for up-front doxycycline treatment. Patients who failed after doxycycline therapy were successfully salvaged with CVP chemotherapy and/or radiotherapy without compromising the long-term outcomes.