The curative potential of radiotherapy (RT) alone as initial treatment for patients with Stage I–II lymphocyte-predominant Hodgkin lymphoma (LPHL) has not been defined well.
The curative potential of radiotherapy (RT) alone as initial treatment for patients with Stage I–II lymphocyte-predominant Hodgkin lymphoma (LPHL) has not been defined well.
Two hundred two patients who were treated between 1969 and 1995 were evaluated in a retrospective, multicenter study.
Patient characteristics were as follows: The median age was 31 years, 75% of patients were male, 80% of patients had Ann Arbor Stage I disease, 1% of patients had bulky disease, 3% of patients had B symptoms, 1% of patients had extranodal involvement, and 80% of patients had supradiaphragmatic disease. The RT fields were a full mantle field in 52% of patients, less than a full mantle field in 24% of patients, an inverted-Y field in 17% of patients, less than an inverted-Y field in 3% of patients, and total lymph node irradiation in 3% of patients. The median dose was 36 Gray. The median follow-up was 15 years. The overall survival (OS) rate at 15 years was 83%, and freedom from progression (FFP) was observed in 82% of patients, including 84% of patients with Stage I disease and 73% of patients with Stage II disease. No recurrent LPHL and only 1 patient with non-Hodgkin lymphoma (NHL) were reported after 15 years. Adverse prognostic factors that were identified on multifactor analysis were as follows: for OS, age 45 years or older (P < 0.0005), the presence of B symptoms (P = 0.002), increasing number of sites (P = 0.015); for FFP, increasing number of sites (P = 0.002). No significant difference was found in FFP in a comparison of patients who received elective mediastinal RT with patients who did not receive mediastinal RT (P = 0.11). Causes of death at 15 years were LPHL in 3% of patients, NHL in 2% of patients, in-field malignancy in 2% of patients, in-field cardiac/respiratory in 4% of patients, and other in 6% of patients.
The current data suggested that RT potentially may be curative for patients with Stage I–II LPHL and raise the possibility that limited-field RT may be used without loss of treatment efficacy. Involved-field RT warrants further investigation for patients with early-stage LPHL. Cancer 2005. © 2005 American Cancer Society.
Lymphocyte-predominant Hodgkin lymphoma (LPHL) accounts for approximately 5% of Hodgkin lymphoma and differs from classic Hodgkin lymphoma in its molecular and clinical characteristics.1–5 LPHL typically presents with localized peripheral lymphadenopathy. Bulky disease, B symptoms, and mediastinal or abdominal involvement are uncommon.6 Traditionally, early-stage LPHL and classic Hodgkin lymphoma have been treated with wide-field radiotherapy (RT) or, more recently, with limited chemotherapy and involved-field RT.7 However, the indolent natural history of LPHL and concerns about late treatment-related toxicity have led to interest in minimizing initial treatment.2, 8–10 Several investigators have advocated strategies used for early-stage, indolent non-Hodgkin lymphoma (NHL), including involved-field RT, or observation if all known disease has been excised.2, 11
To our knowledge to date, few large studies with long-term follow-up have been reported for patients with LPHL who were treated with limited RT alone.2, 6, 8, 9, 12–16 The objectives of the current study were to examine the long-term results of initial RT alone, to identify favorable patient subsets, and to explore the effect of reducing RT field size on treatment efficacy.
This was a multicenter retrospective study of RT alone for patients with LPHL. Investigators identified patients by review of institutional data bases. Eligible patients had a classification of LPHL prior to initial treatment, Stage I or II (clinical or pathologic) disease, initial treatment by RT alone between 1969 and 1995 (inclusive), and adequate documentation of disease characteristics, staging investigations, and treatment. Some patients have been included in previous series.17–20 Staging was according to the Ann Arbor system.21 If involvement of a site was ambiguous, then it was coded as uninvolved, uncertain stage was coded as the lower of two possible stages, and B symptoms were coded as present only if they were unequivocal on chart review. The current analysis was based on the pathologic classification at the time of original treatment. Where pathology was reviewed in a hospital pathology department prior to treatment, the review classification has been used. Additional review of pathologic material has not been performed as part of the current study. RT dose and sites irradiated and details of salvage treatment were recorded. RT fields were classified as either a full mantle field, less than a full mantle field (if any component of the mantle was omitted), an inverted-Y field, less than an inverted-Y field, or total/subtotal lymph node fields. In this retrospective study, reasons for selection of field size and the quality of RT delivery could not be determined. Data regarding the date and sites of recurrent LPHL, the subsequent occurrence of NHL, other malignancies, and the date and cause of death were obtained from departmental records, managing physicians, and patients. It was not possible to distinguish recurrent LPHL reliably from another NHL in every patient, because these events were not subject to pathology review for the current study. Hence, recurrent LPHL and any subsequent NHL (other than unequivocal T-cell NHL) were grouped together for analysis of the time to disease progression. Data from case record forms were entered into a study-specific Microsoft Access data base.
The close-out date for the analysis was September 4, 2001. All patients were followed up to this date unless they were lost to follow-up. Two-sided P values are reported throughout, with no formal adjustment for multiple comparisons. Statistical analyses were performed using S-PLUS and SPSS statistical software.22, 23
Overall survival was measured from the commencement of RT (unless otherwise stated) to the date of death from any cause. Disease progression was defined as either the recurrence of LPHL or the development of any NHL (other than T-cell NHL). Freedom from progression was measured from the commencement of RT to the date of recurrence of LPHL or the development of any NHL (other than T-cell NHL). Freedom from second recurrence of LPHL was measured from the commencement of first salvage treatment to the date of second recurrence of LPHL. The Kaplan–Meier product-limit method was used, and the standard error (SE) of each survival rate was calculated using the Greenwood formula.
Unifactor analyses were performed to examine the significance of individual factors for overall survival and freedom from progression (factors examined are shown in Table 1). The Mantel–Cox log-rank test was used to compare survival between subgroups, and a test for trend was used when there were more than two subgroups consisting of ordered categories. An exact log-rank test based on the distribution of the sum of independent, hypergeometric, random variables was used with binary factors in the presence of small numbers. Results were considered statistically significant if the P value was < 0.05.
|Category||Overall survival||Freedom from progression|
|No. of patients||No. of deaths||Rate at 15 yrs (SE)||P value||No. of progressions||Rate at 15 yrs (SE)||P value|
|All patients||202||41||83 (3)||30||82 (3)|
|< 45 yrs||145||15||94 (2)||22||83 (4)|
|≥ 45 yrs||57||26||55 (8)||< 0.0001||8||82 (6)||0.63|
|Male||151||26||85 (3)||21||83 (4)|
|Female||51||15||75 (7)||0.044||9||79 (6)||0.39|
|Stage I||162||29||86 (3)||20||84 (4)|
|Stage II||40||12||70 (8)||0.16||10||73 (7)||0.031|
|No||195||37||84 (3)||30||82 (3)|
|No||199||40||83 (3)||29||82 (3)|
|No||200||41||82 (3)||30||82 (3)|
|No. of sites with disease involvement|
|One site||162||29||86 (3)||20||84 (4)|
|Two sites||27||7||76 (9)||4||85 (7)|
|Three sites||9||2||71 (18)||4||53 (17)|
|More than three sites||4||3||25 (22)||0.032||2||33 (27)||0.0001|
|Supradiaphragmatic||161||31||84 (3)||24||82 (4)|
|Infradiaphragmatic||41||10||77 (7)||0.39||6||82 (7)||0.92|
|No||139||30||78 (4)||21||81 (4)|
|Yes||63||11||91 (4)||0.09||9||85 (5)||0.39|
|CT chest and abdomen performed|
|None performed||74||23||84 (4)||13||83 (5)|
|Only 1 performed||25||4||82 (8)||3||87 (7)|
|Both performed||102||14||82 (5)||0.84||14||76 (7)||0.85|
|Pathology review performed|
|No||82||16||82 (5)||12||80 (6)|
|Yes||120||25||83 (4)||0.87||18||83 (4)||0.98|
|1969–1979||52||16||86 (5)||10||81 (6)|
|1980–1989||95||19||82 (4)||16||81 (4)|
|A||102||23||80 (4)||18||80 (4)|
|B||30||6||82 (7)||2||96 (4)|
|C||24||7||80 (9)||3||87 (7)|
|D||22||4||90 (7)||2||89 (7)|
|E-I||24||1||94 (6)||0.56||5||47 (21)||0.32|
To identify prognostic factors that independently predicted overall survival and freedom from progression, multifactor analyses were performed. Factors that were included in the multifactor analyses were the same as those used in the unifactor analyses, apart from treatment center, and the exclusion of extranodal sites for the analysis of overall survival, and extranodal sites and B symptoms for the analysis of freedom from progression (because no relevant events occurred in the patients with these characteristics). The Cox proportional-hazards model with a stepwise, backward regression procedure was used to identify significant factors (P values for entry and removal of 0.01 and 0.05, respectively, were used). The likelihood-ratio test was used to test for the significance of a factor after adjusting for other factors already in the model. The prognostic significance of each factor identified was summarized using the hazard ratio that represented the relative death rate or disease progression rate for a given group, relative to a baseline group, together with its 95% confidence interval.
Cumulative incidence was estimated using competing risk analyses for 1) site of first recurrence of LPHL (in-field with or without out-of-field, out-of-field only, recurrence site unknown, and death without recurrence), 2) second malignancy (solid tumor in-field, solid tumor out-of-field, NHL, leukemia, and death), and 3) cause of death (LPHL, NHL, potentially RT-related in-field toxicity [malignancy, cardiac, respiratory], complications of salvage treatment, unrelated cause, and cause unknown).24
There were 202 patients from 9 centers (Peter MacCallum Cancer Centre, Victoria; Royal Brisbane Hospital, Queensland; Royal Prince Alfred Hospital, New South Wales; Royal Adelaide Hospital, South Australia; St. Vincent's Hospital, New South Wales; East Coast Cancer Centre, Queensland; Sir Charles Gairdner Hospital, Western Australia; Alfred Hospital, Victoria; and Newcastle Mater Misericordiae Hospital, New South Wales). Patient age was between 2 years and 79 years (median age, 31 yrs). The estimated median follow-up duration was 15 years.25 Thirty-three patients were lost to follow-up; of these, 26 patients had been followed for at least 5 years and 16 patients had been followed for at least 10 years. Patient characteristics are presented in Table 1.
All patients were considered to have LPHL based on original pathology reports or pathology review at the treating institution, which was performed in 120 patients. Fourteen patients who initially received another diagnosis (nodular sclerosis in 3 patients, mixed cellularity in 4 patients, follicular lymphoma in 2 patients, and unclassified in 5 patients) were reclassified with LPHL at review prior to treatment. Staging investigations included laparotomy in 31% of patients, splenectomy in 29% of patients, abdominal imaging in 92% of patients (lymphangiogram in 31% of patients, computed tomography scan in 39% of patients, and both in 22% of patients), chest computed tomography scan in 52% of patients, gallium scan in 23% of patients, and bone marrow examination 84% of patients. RT fields were a full mantle field in 52% of patients, less than a mantle field in 24% of patients, a full inverted-Y field in 17% of patients, less than a full inverted-Y field in 3% of patients, and total lymph node irradiation (with or without the spleen) in 3% of patients. The median RT dose was 36 Gray (Gy) (30–34 Gy in 12 patients, 35–36 Gy in 92 patients, 37–40 Gy in 89 patients, and 41.0–50.4 Gy in 9 patients).
The estimated percentage of patients who remained alive was 88% (SE, 2%) at 10 years, 83% (SE, 3%) at 15 years, and 74% (SE, 4%) at 20 years. (Fig. 1) The results of a unifactor analysis for potential prognostic factors are shown in Table 1. After the multifactor analysis, prognostic factors that remained significant were patient age (P < 0.0005), B symptoms (P = 0.002), and the number of involved sites (P = 0.015) (Table 2).
|Significant variable||HR||95% CI||P value|
|< 45 yrs||1.0|
|≥ 45 yrs||11.5||5.3–24.8||< 0.0005|
|No. of involved sites per categoryb||1.8||1.2–2.9||0.015|
|Freedom from progression|
|No. of involved sites per categoryb||2.1||1.4–3.1||0.002|
Thirty patients were diagnosed with disease progression (25 patients with LPHL and 5 patients with NHL). The estimated percentage of patients who were progression free was 88% (SE, 2%) at 10 years, 82% (SE, 3%) at 15 years, and 82% (SE, 3%) at 20 years (Fig. 1). The results of unifactor analyses for potential prognostic factors are shown in Table 1. On multifactor analysis for freedom from progression, the only significant factor identified was the number of involved sites (P = 0.002) (Table 2). Freedom from progression by the number of involved sites is illustrated in Figure 2.
A subset analysis was performed on the 146 patients who had supradiaphragmatic disease without mediastinal involvement and were treated with supradiaphragmatic RT fields only. Mediastinal fields were omitted for 44 patients. Inclusion of the mediastinal field was associated with a 15-year progression-free rate of 79% (SE, 5%) compared with 92% (SE, 5%) when the mediastinum was omitted from the field (P = 0.11) (Fig. 3). Patients who had mediastinal irradiation had a relative risk of progression 2.6 times (95% confidence interval, 0.8–9.0 times) that of patients who did not have mediastinal irradiation. After adjusting for the number of involved sites (the only factor identified from the multifactor analysis of freedom from progression), the result of the comparison between the 2 subgroups remained statistically insignificant (P = 0.12). A further analysis, which also adjusted for the disease location (high neck vs. other), did not alter this finding (P = 0.18).
Information on sites of progression was available in 24 of 25 patients who had recurrent LPHL recorded. Four patients had in-field progression (with or without distant sites) in sites that received 36–40 Gy. There were no in-field recurrences in 35 patients who received 30–35 Gy. Seventeen patients failed out-of-field only in lymph nodes with or without the spleen, with 14 of these progressions occurring solely on the opposite side of the diaphragm to the initial presentation. Disease progression occurred in the spleen in six patients, in the liver in three patients, and in bone marrow in two patients (two patients had more than one site). No patient had an isolated mediastinal progression. Cumulative incidences of progression at 15 years were in-field progression with or without out-of-field progression in 2% of patients (SE, 1%) and out-of-field progression only in 12% of patients (SE, 3%).
The analysis of salvage treatment was based on the 25 patients who had recorded recurrences of LPHL. The estimated overall survival at 10 years after the first recurrence of LPHL was 60% (SE, 11%) (Fig. 4A). After the first recurrence of LPHL, 22 patients had salvage treatment (chemotherapy in 13 patients, RT in 5 patients, and both in 4 patients). The estimated percentage of patients who were free from second recurrence of LPHL at 10 years following salvage treatment was 52% (SE 12%) (Fig. 4B). The following numbers of patients were free from second recurrence of LPHL according to salvage regimen: five of eight patients who received mechlorethamine, vincristine, procarbazine, and prednisone (MOPP)-like regimens; one of three patients who received anthracycline-based regimens; four of five patients who received MOPP-doxorubicin, bleomycin, vinblastine, and dacarbazine-like regimens; none of one patient who underwent transplantation; and six of nine patients who received RT-containing regimens.
Reported second malignancies were NHL in 9 patients, acute leukemia in 1 patient, and other in 18 patients. Nine second malignancies were in-field carcinomas (breast in two patients, lung in one patient, bowel in one patient, gastrointestinal in two patients, cutaneous melanoma in one patient, and unknown primary in two patients), and nine were out-of-field carcinomas. Cumulative incidences of second malignancy at 15 years were all malignancies, 11% of patients (SE, 3%); carcinoma in-field, 4% of patients (SE, 2%); carcinoma out-of-field, 3% of patients (SE, 1%); NHL, 4% of patients (SE, 1%); and leukemia, 1% of patients (SE, 1%).
Of 41 deaths reported, 8 deaths were due to lymphoma (LPHL in 5 patients and NHL in 3 patients). Seventeen deaths potentially were related to initial RT (in-field malignancy in 8 patients and cardiac or respiratory disease after mediastinal RT in 9 patients). One death was a possible late complication of salvage RT. Thirteen deaths were considered unrelated to LPHL or treatment. The cause of death was unknown for two patients. Cumulative incidences of cause of death at 15 years are shown in Table 3.
|Cause of death||Estimated cumulative incidence ± SE (%)|
|LPHL||2.9 ± 1.3|
|NHL||1.5 ± 1.1|
|Possible complication of initial treatment||6.2 ± 2.0|
|In-field malignancy||2.2 ± 1.1|
|Cardiaca||3.0 ± 1.5|
|Respiratorya||1.0 ± 0.7|
|Complication of salvage||0.5 ± 0.5|
|Disease or treatment unrelated||5.7 ± 1.8|
|Unknown||0.6 ± 0.6|
With long-term follow-up in a large study cohort, we found that approximately 80% of patients were free from disease progression 15–20 years after treatment with RT alone for early-stage LPHL. Although previous studies of early-stage LPHL have varied with respect to RT field size, chemotherapy use, and the length of follow-up, most have reported comparable 7–12-year freedom from progression rates of 70–80%.2, 6, 9, 12, 13, 15 The largest study was from the European Task Force and included 256 patients who were classified previously with Stage I–II LPHL.2 Among those patients with LPHL, on pathology review, the 8-year failure-free survival rate was 85% for patients with Stage I disease and 71% for patients with Stage II disease. Patient characteristics in the current study were similar to those in previous reports, with a preponderance of males, Stage I disease, peripheral disease sites, and rarity of bulk and B symptoms.2, 6, 15, 16 In our multifactor analysis, the only prognostic factor for progression was the number of involved sites. Patients with 1 or 2 disease sites had a significantly greater chance of long-term freedom from progression after RT than patients with ≥ 3 sites (P < 0.0001) (Fig. 2). Approximately 50% of the patients who developed recurrences after RT experienced durable second remissions after salvage treatment.
Some investigators have reported a continuing pattern of recurrence after treatment for LPHL, raising questions concerning the curability of this condition.2, 8 In the Task Force report, there were 8 failures between 10 years and 15 years and 4 failures after 15 years.2 In a series with a median follow-up of 76 months, Orlandi et al. described a continuing pattern of recurrence, with many of the 22 incidents of progression occurring 5–10 years after treatment, and 2 incidents occurring after 10 years.8 In a report by Regula et al., there were 2 recurrences after 10 years; and, in a study by Borg-Grech et al., there were 4 recurrences reported after 10 years.9, 13 However, these data should be interpreted with caution. For example, in the Task Force report, it is unclear how many late recurrences occurred in patients with Stage III–V disease or in patients who were treated with initial chemotherapy alone.2 In the study by Orlandi et al., the long-term remission rate of 45% at 10 years was lower than that reported in most other series. It is interesting to note that 25% of their patients had Stage III–IV disease, 15% had B symptoms, and 28% had nonperipheral disease sites, which are uncharacteristic of LPHL.8 In the current study, 31 patients were on follow-up and were free from progression at 20 years after initial treatment. It also is interesting to note that no recurrences of LPHL were reported after 15 years, and only 1 recurrence of NHL was reported after 15 years, suggesting that limited RT potentially is curative for a proportion of patients with LPHL. This finding is supported by previous studies that did not report frequent late recurrences;, although, in some studies, follow-up was limited.14–16, 26
The results of RT treatment for LPHL appear to differ from those reported for early-stage classic Hodgkin lymphoma or follicular lymphoma. In a previous study of mantle-field RT for patients with clinical Stage I and II Hodgkin lymphoma, we reported that patients diagnosed with LPHL had a 10-year freedom from progression rate of 80% compared with 50% for patients diagnosed with classic Hodgkin lymphoma, a difference that was significant after adjusting for other prognostic factors.17 Studies of early-stage follicular NHL treated with RT alone have reported durable remission rates of 40–50%.27, 28 These results suggest that LPHL has a natural history that is distinct from both classic Hodgkin lymphoma and follicular NHL.
The current results were obtained using RT that was confined to one side of the diaphragm in all but six patients. Limited-field RT may be adequate due to the low incidence of occult abdominal or mediastinal involvement in LPHL.6 Similar to other studies, we found that deaths due to LPHL were outnumbered by deaths from other malignancies and cardiorespiratory causes, some of which may have been attributable to initial treatment.2 For example some malignancies and cardiac events may have been related to mediastinal irradiation. Hence, an important question is whether the RT volume can be reduced without loss of treatment efficacy. We evaluated a subgroup of patients who had supradiaphragmatic disease but no mediastinal involvement and compared the outcomes in patients who did and did not receive elective mediastinal irradiation. There was no statistically significant difference in freedom from progression between the two subgroups. Because of the possible confounding effect of patient selection, we adjusted the results for the number of involved sites (the only independent factor for freedom from disease progression) and still could identify no adverse effect of omitting mediastinal fields. However, we cannot exclude possible bias resulting from unknown selection criteria that determined the RT field size, and the small sample size limited the statistical power of this comparison. Given these caveats, the data raise the possibility that RT fields may be limited to volumes smaller than a classic mantle field without loss of efficacy. This was supported by the pattern-of-failure analysis, which identified no incidents of isolated mediastinal recurrence. A similar finding was reported by Schlembach et al.26 Like the treatment volume, the minimal effective dose of RT for LPHL is unknown. Our very low in-field failure rate suggests that 35–36 Gy may be adequate to control LPHL. The lack of in-field recurrences in 12 patients who received 30–34 Gy suggests that these lower doses may be sufficient, but this needs confirmation in additional series. For the small number of recurrences that occurred after initial RT, salvage treatment appeared to produce durable second remissions, suggesting that a proportion of patients with recurrent LPHL can be cured with conventional therapy.
Chemotherapy is used routinely in the management of early-stage classic Hodgkin lymphoma, and it has been shown that rituximab produces complete response rates of 41–57% for LPHL.29, 30 However the contribution of systemic therapy to the initial management of LPHL is uncertain. One small, nonrandomized, retrospective study compared patients who received RT alone for LPHL with patients who received both chemotherapy and RT and found no significant difference in outcomes.12 Any benefit from the addition of systemic treatment to RT is likely to be small and, given the rarity of LPHL, is unlikely to be resolved in a prospective, randomized trial.
An alternative management strategy, observation alone, has been advocated for patients who have undergone complete resection of their disease at biopsy.2, 11 The rationale proposed for observation alone is that LPHL has an indolent natural history; is associated with a very low disease-specific mortality; and, in any event, may not be curable, as reflected in the late recurrences sometimes reported. These observations, combined with concerns about potential treatment-related toxicity, have led to the observation of some patients after initial biopsy. A potential disadvantage of observation is the risk that patients ultimately may develop recurrences with a less favorable disease presentation and a diminished potential for cure. Furthermore, it is possible that the low mortality generally attributed to LPHL is due to the efficacy of initial treatment and may not be seen after a watch-and-wait policy is used.
Because not all pathology reports were reviewed, it is possible that a proportion of patients in the current series may not be classified with LPHL using modern diagnostic criteria. However, several lines of evidence suggest that the current data are likely to be representative of the results that would be achieved in a pure cohort of patients with LPHL. First, our patient characteristics are very similar to those of the cohort reviewed in the Task Force study. Second, in the Task Force series, patients with lymphocyte-rich classic Hodgkin lymphoma had outcomes similar to those achieved in patients with LPHL.2 Finally, in a U.K. study, pathology from 90 patients initially classified as LPHL was reviewed and was reclassified as mixed cellularity (23%), nodular sclerosis (11%), and NHL (10%). The remission duration for patients who were not considered to have LPHL was significantly lower than for patients who had LPHL confirmed on review.14 These results are consistent with published results in patients with classic Hodgkin lymphoma and follicular NHL who were treated with limited RT alone.17, 27 Thus even if the current series included a proportion of patients who currently would be classified with classic Hodgkin lymphoma or follicular NHL, it is unlikely that we overestimated the results of RT for LPHL. Another potential limitation of this study was the difficulty in distinguishing reliably between the progression of LPHL and the development of a new NHL. However, because we included both recurrent LPHL and the development of NHL in our analysis of freedom from progression, the reported results may represent a conservative estimate of treatment efficacy.
The large size and long follow-up of the current study allow meaningful observations about long-term outcome, prognostic factors, and whether a plateau on the progression-free curve was attained. The current study data suggest that RT potentially is curative for patients with early-stage LPHL, with excellent results achieved in patients with Stage I or Stage II disease who have 2 sites of involvement. An evaluation of the impact of RT field size, adjusting for significant prognostic factors, suggested that a full mantle field may not be required to maintain excellent freedom from progression. It is possible that comparable results may be obtained with moderate-dose, involved-field RT, which may lead to a reduction in late toxicity. The long-term results of involved-field RT, initial observation, or novel systemic therapies for patients early-stage LPHL will require further prospective study.
The authors thank Dr. John Seymour for his helpful comments on the article, Vicki Walcher for data management, and Cathy Faull for secretarial assistance.