Second malignancies after treatment of classic Hodgkin lymphoma

INTRODUCTION — Most patients with classic Hodgkin lymphoma (cHL) achieve long-term survival without a recurrence of cHL, but late complications of treatment have emerged as competing causes of death and morbidity in survivors. Late complications of treatment for cHL include cardiovascular disease, pulmonary disease, and second malignancies.

Survivors of cHL treatment have an increased relative risk for developing a second cancer compared with the general population, but the absolute incidence of these cancers varies widely. The types of second cancer (eg, solid tumor, acute leukemia, another lymphoma) and the time course of development vary with the previous treatment (ie, chemotherapy and/or radiation therapy) and host factors (eg, age at treatment, smoking).

This topic discusses second malignancies that occur in patients who have been treated for cHL.

Long-term follow-up of patients with HL, risk of second malignancy after hematopoietic cell transplantation, and other complications are presented separately. (See "Approach to the adult survivor of classic Hodgkin lymphoma" and "Bleomycin-induced lung injury" and "Secondary cancers after hematopoietic cell transplantation" and "Clinical manifestations, diagnosis, and treatment of anthracycline-induced cardiotoxicity" and "Risk and prevention of anthracycline cardiotoxicity".)

GENERAL OBSERVATIONS — Patients treated with chemotherapy and/or radiation therapy (RT) for cHL have an increased risk of developing solid tumors and hematologic malignancies when compared with the general population [1-4]. These second cancers have a significant impact on survival from HL. The relative risk of developing a second malignancy has varied greatly among studies and depends on factors related to the patient and the treatment, including:

●Radiation dose and fields

●Chemotherapy agents and doses administered

●Patient age at treatment

●Length of time after treatment

●Family history

●Smoking history

In addition to these factors, certain patients may have a genetic predisposition toward developing second malignancies after treatment of cHL. As an example, a genome-wide association study in survivors of cHL treated with RT reported that patients with certain single nucleotide polymorphisms at chromosome 6q21 were more likely to develop a second malignancy [5]. Further study of candidate genes in this region may reveal new data regarding the pathogenesis of second malignancies in this population.

Effect of elapsed time — Most deaths in the first 5 to 10 years after treatment are associated with relapse of HL, while later deaths are more likely to be due to other causes (eg, second malignancy, cardiovascular disease). The cumulative incidence of mortality from HL is exceeded by mortality due to other causes after 14 years from treatment (figure 1) [6,7]. Although death from HL abates with time, the absolute excess risk of mortality from other causes remains elevated, with no apparent diminution of the excess risk beyond 20 to 30 years [4,6,8-10]. Importantly, all-cause mortality risk among long-term cHL survivors is significantly lower in patients treated in a more recent era [11].

Different types of cancer exhibit distinct temporal patterns of development after treatment of cHL. As an example, in patients who received alkylator-based regimens, the risk of developing leukemia peaks at five to nine years post-treatment, whereas solid tumors typically develop after a latency period of ≥10 years [12]. In one series, the median time to onset of acute leukemia, non-Hodgkin lymphoma (NHL), and solid tumors was 5, 7.3, and 12.2 years, respectively [13]. The relative and absolute excess risks for particular cancers in an individual may differ from those observed in the entire patient population.

Relative and absolute risks — Solid tumors account for most second malignancies in cHL survivors, with breast, lung, and gastrointestinal cancer being the most common. Although hematologic cancers are less common overall, the relative risk compared with the general population is higher.

The estimated overall relative risks in 11 large cohort studies were:

●Leukemia – 10 to 80-fold

●NHL – 3- to 35-fold

●Solid tumors (lung, breast, bone, stomach, colon, thyroid, melanoma) – >2-fold

Leukemia and NHL are diseases with a low incidence in the general population; as a result, an increase in their relative risk following the treatment of cHL (ie, up to 80-fold) does not translate into a high likelihood of developing leukemia or NHL. In addition, the risk of leukemia in this study reflects the use of chemotherapy regimens, such as MOPP and MOPP/ABVD, which are no longer the standard of care. On the other hand, even though the relative risk of developing a solid tumor is lower (ie, twofold), the cumulative number of solid tumors that develop over the entire follow-up period is much higher than the number of cases of leukemia or NHL, because the incidence of solid tumors in the general population is so much higher (table 1) [14].

The absolute number of second cancers in patients with cHL that may be attributable to treatment (ie, excess cases) is estimated to be 40 to 90 malignancies per 10,000 patients per year (0.4 to 0.7 percent per year) (table 1) [8,12-17]. Several retrospective case series have reported an actuarial risk of second cancer of 11 to 18 percent at 15 years [15,16]; 20 percent at 20 years [18,19]; 28 percent at 25 years [14]; and 33 percent at 30 years [4].

A combined analysis of three large studies including 6292 patients showed that ≥10 years after first treatment, lung cancer contributed most to the absolute excess cancer risk, with 34 excess cases per 10,000 cHL patients per year [12,15,18]. NHL (28/10,000 per year) was second, and leukemia (10/10,000 per year) third. In females, breast cancer accounted for the largest absolute excess risk (40/10,000 per year) in 10-year survivors.

Treatment risk factors — The risk for various cancers varies with the specific treatment. As an example, chemotherapy that includes alkylating agents (eg, procarbazine, cyclophosphamide), such as escalated BEACOPP, second-line chemotherapy, and autologous hematopoietic stem cell transplant (HCT), are at increased risk for therapy-related myeloid neoplasms (t-MN), including myelodysplastic syndromes (MDS) and acute myeloid leukemia (AML), whereas ABVD, which does not include alkylating agents is not associated with the development of t-MN. The development of secondary lung cancer appears to be related to the use of radiotherapy and/or chemotherapy.

A meta-analysis of randomized studies of cHL treated from 1984 to 2007 identified 9498 patients with individual patient-level data [20]. With a median follow-up of more than seven years, intensified chemotherapy was associated with improved progression-free survival (PFS), but a higher risk of t-MN, including MDS and AML. In patients who received combined modality therapy (CMT) versus chemotherapy alone, there were no differences in PFS or overall survival (OS), but there were more secondary malignancies in patients who received RT. In comparing involved-field versus extended-field and lower dose (20 Gy) versus higher dose RT, there was no significant change in rates of secondary malignancies.

A series of 5798 patients with cHL treated in Britain between 1963 and 2001, the relative risk (RR) of developing a second malignancy was higher with combined chemotherapy plus RT (RR 3.9) than with chemotherapy alone (RR 2.0) [21]. Among patients treated with chemotherapy alone, the excess second malignancies were leukemia, NHL, and lung cancer. In this population, the excess risk peaked five to nine years after chemotherapy administration and was close to baseline by 15 years. In contrast, combination chemotherapy plus RT was associated with several other types of second malignancy and the excess risk remained for 25 years or longer.

Prognosis — Patient survival after the diagnosis of a second cancer is generally poor. In our series, the five-year survival rate after the development of a second malignancy was 38 percent, with the worst prognosis seen in those developing acute leukemia or lung cancer [8]. Patients who develop acute leukemia or MDS have a two-year survival rate of approximately 10 percent, while those who develop NHL or a solid tumor have 5- and 10-year survival rates of approximately 32 and 23 percent, respectively [16]. Similarly, in another study, the median survival from diagnosis of second tumor was approximately 42 months and was especially poor in patients developing a second cancer following an HL diagnosis at the age of 40 years or older (median survival: approximately one year) [13].

Second cancers have a significant impact on survival from cHL [14,16,22]. The range of findings can be illustrated by the following observations:

●In a 20-year follow-up study of almost 1300 patients (median follow-up 9.2 years) who were treated between 1966 and 1986, the cumulative risk of dying from cHL or acute treatment toxicity was 33 percent, while the risk of dying from a second cancer was 14 percent and that from all other causes was 20 percent [18].

●Virtually identical findings were noted in a report from the International Database on Hodgkin's Disease [16]. The overall probability of 20-year survival in 14,315 patients with HL who were treated between 1960 and 1987 (average follow-up 6.7 years) was 50 percent; 37 percent died from cHL or acute treatment toxicity, and 10 percent died from a second cancer.

Patients who develop a second cancer are also at risk of developing a third cancer. In one series of 3122 Dutch patients treated for cHL between 1965 and 1995 and followed for a median of 23 years, 832 developed a second malignancy and 126 developed a third [23]. The median time from cHL to second malignancy was 19 years and from second to third was four years. Cancer screening in long-term survivors of cHL is discussed separately. (See "Approach to the adult survivor of classic Hodgkin lymphoma", section on 'Late complications'.)

NON-HODGKIN LYMPHOMA — While the relative risk (compared with the general population) of non-Hodgkin lymphoma (NHL) is increased among cHL survivors, the absolute number of cases is low due to the low incidence in the general population. The reported incidence of NHL in large series of cHL survivors has ranged from 0.9 percent at an average follow-up of 6.7 years (0.6 percent in the subset in continuous complete remission) [16,24] to 1.6 percent at 15 years [15]. The relative risk for NHL progressively increased with follow-up time in most [14-16,25] but not all studies [12,26]. Diffuse large B cell lymphoma is the most common subtype, accounting for 78 percent of the cases [24].

The largest experience comes from the International Database on Hodgkin's Disease, which evaluated the course of 12,411 patients with an average follow-up of 6.7 years [16]. One-hundred-six cases of NHL were observed with a continuous increase in relative risk with follow-up in males (from 28.2 during the five- to nine-year period, to 120 during the 15- to 19-year period). In females, the risk increase was confined to the first 15 years, with a peak relative risk of 45.4 during the five- to nine-year observation period.

Risk factors — There is little agreement on definitive risk factors for NHL in patients treated for cHL. It is possible that differences among studies represent differences in the patient populations examined. It is also possible that pathologic misclassification may confound the results, with recurrent cHL being misdiagnosed as NHL or NHL being initially misdiagnosed as cHL. This issue is of lesser concern when modern histopathologic techniques are used. Only a few studies have included expert review of the original tissue diagnosis as part of the study [12,18,24]. The risk of NHL appears to have declined over the past decade, perhaps related to the use of less immunosuppressive chemotherapy regimens in the initial treatment of cHL. In our experience, second NHLs seen after modern treatment regimens for cHL are more likely to be low-grade subtypes; a greater frequency of high-grade NHL was seen previously when more immunosuppressive regimens were used.

Treatment modality — There are conflicting data concerning the impact of initial treatment modality on the risk of subsequent NHL. Several studies have found no apparent differences in increased risk between patients treated with chemotherapy alone, radiation therapy (RT) alone, or a combination modality [12,13,15,26]. However, other reports suggested an increased risk with combined modality treatment and treatment for relapse. A large International Database study included 106 patients who developed NHL; combined modality treatment and any type of treatment for relapse were associated with increased risk for NHL with relative risks of 2.2 to 4.5 [16].

Host factors — There are also conflicting data on the importance of host factors such as sex and age on the risk for NHL. Some studies noted no relationship between these factors and NHL [27], while the International Database and report from the Netherlands Cancer Institute found that increasing age over 30 and, to a lesser degree, male sex were risk factors [16,18].

Tumor factors — The development of NHL may be related in part to the diagnosis of cHL itself rather than its treatment. In particular, the histologic subtype of cHL and the immunosuppressive effects of cHL may influence the risk of NHL.

The risk of developing NHL differs among patients with different cHL histologies, with the greatest risk for NHL occurring in patients with nodular lymphocyte predominant disease (NLPHL) and the lowest risk in patients with nodular sclerosis HL (NSHL) [16,18,27,28]. In one study, for example, the relative risk of NHL among cHL survivors compared with the general population was 55.6 with NLPHL compared with 9.8 with NSHL [27]. In another report, NHL occurred in 22 of 3033 patients with cHL randomized into clinical trials by the British National Lymphoma Investigation [28]. The frequency of NHL ranged from 3.8 percent in NLPHL to 0.3 percent in NSHL. It has been suggested that the transformation to NHL may be part of the natural history of NLPHL, rather than being related to therapy [28]. Alternatively, it is often difficult to distinguish NLPHL from NHL, and in some cases the "secondary NHL" may simply be a recurrence of the initial disease.

In addition, cHL may be associated with a general state of immunosuppression [28,29], which can predispose to the development of NHL as in transplant recipients. Support for this hypothesis comes from a detailed study of 14 patients who developed NHL within a median of 136 months from the diagnosis of cHL [29]. The high prevalence of extranodal sites, either intermediate or high-grade lymphoma subtype, and immunophenotypic findings typical of NHL of B cell lineage patients were similar to those of NHL arising in immunosuppressed patients. It has been our experience that the incidence of NHL after cHL has decreased, perhaps due to the reduction or disuse of myelosuppressive agent combinations such as MOPP. In addition, we are now seeing more indolent NHL after cHL. (See "Treatment and prevention of post-transplant lymphoproliferative disorders".)

Treatment and prognosis — There are limited data concerning the prognosis of NHL following the treatment of cHL. Initial studies reported poor outcomes [30]. However, later studies suggest that some of these patients may be curable with aggressive multiagent chemotherapy [13,16,24,29].

●Of the 106 patients with secondary NHL in the International Database on Hodgkin's Disease (diagnosed between 1960 and 1987), 45 were alive at the time of the report with a five-year probability of survival of 30 percent [16].

●A retrospective analysis of 52 patients who developed NHL after treatment for cHL (diagnosed from 1981 to 1998) reported a two-year overall survival rate of 30 percent [24].

●A retrospective analysis of patients in the Surveillance, Epidemiology, and End Results (SEER) program database treated from 2000 to 2004 reported 5- and 10-year survival rates for patients with secondary NHL of 54 and 41 percent, respectively [31].

SOLID TUMORS

Aggregate risks — Solid tumors account for approximately one-half to two-thirds of second malignancies developing after 15 or more years of follow-up in patients treated for cHL [4,12-14,18,32]. The risk of solid tumor development depends greatly on the type of treatment, age at treatment, follow-up interval, and family history.

The excess risk of lung and breast cancers appears to be related to their proximity to the radiation portals. As an example, in one study of 202 patients surviving at least five years after treatment for cHL, 20 of the 26 solid tumors following primary treatment were found within or adjoining the irradiated fields [19].

The overall relative risk for solid cancer, compared with general population expectations, is approximately 1.5 to 4.5 in long-term follow-up of studies of predominantly adults (table 1) [13,15,16,18,33-38] but much larger (approximately 12) in children [39,40]. The cumulative risk of solid cancer in adults is approximately 9 to 13 percent at 15 to 20 years (table 1) [12,14-16,18,36,37,41-43] and 18 to 33 percent at 30 years [4,33]. In patients treated as children, the reported cumulative risk has been 4 percent at 15 years [39,40] and 27 percent at 30 years [40,44].

There is a small increase in risk in the first five years, but this risk continues to increase out to at least 20 to 30 years [8,13,15,16,18,39], and perhaps indefinitely [33]. The risk is inversely related to age at initial treatment. In one study, compared with the general population, patients who were first treated at age 31 to 39, 21 to 30, and 20 years of age or less, had a solid tumor relative risk of 4.9, 6.9, and 12.7, respectively [14].

The risk of solid cancer overall is increased after radiation therapy (RT) [12,14,15,34,39,42,45]. Some studies have found an increased risk for solid cancers overall in patients treated with chemotherapy [12,37,42,43,46,47], while others have not [15,18,36,48]. The data supporting the role of chemotherapy as a risk factor for secondary solid tumors are most compelling for non-Hodgkin lymphomas and lung cancer [49]. (See 'Non-Hodgkin lymphoma' above and 'Lung cancer' below.)

Family history confers an increased risk of cancer, as it does in the general population. The risk of developing colorectal, lung, or breast cancer was increased at least two- to threefold in cHL survivors with a first-degree relative with that site-specific cancer, compared with cHL survivors without a family history of those solid cancers [38]. Survival of patients with a second cancer after cHL treatment was not worse in those with an affected first-degree relative compared with those without a positive family history.

Large cohort studies of patients with cHL have shown increased rates of a number of solid tumors including lung cancer, breast cancer, gastrointestinal cancer, melanoma and nonmelanoma skin cancers, bone and soft tissue cancers, and thyroid cancer [12-16,34,36,39,42,50-55]. In the largest studies, an elevated risk of salivary gland cancer was also observed [16,50]. There have also been significantly increased risks in some studies for cancers of the mouth, tongue and pharynx, stomach, small intestine, colon, rectum, pancreas, pleura, cervix, ovary, and urogenital cancers overall [12,13,15,16,18,34,36,39,40,42,50,51,55] and, in children, for cancers of the esophagus and brain [39,40,51].

Lung cancer — Lung and breast cancer are the two most common solid malignancies in most large cohort studies with 15 or more years follow-up [12,14,15,42]. In the British National Lymphoma Investigation, for example, one-third of all excess malignancies were lung cancers [12]. In contrast, an American study found that breast cancer was most common, with only 10 percent of excess second malignancies being lung cancers [13]. The relative risk of lung cancer is only slightly increased in the first five years after treatment, with larger relative risks (usually five or more) thereafter until at least 20 years [12,14-16,36,42,50,56]. The relative risk of lung cancer is highest in patients treated with both chemotherapy and radiation before age 25; after this age, the relative risk of second lung cancers decreases in patients treated with radiation but not with chemotherapy [40,49,56]. Lung cancers after chemotherapy occur earlier than those following RT.

The prognosis of patients diagnosed with lung cancer following treatment of cHL is worse than that of patients with de novo lung cancer. This was illustrated in a retrospective analysis that compared outcomes among the 187 of 22,648 cHL survivors who developed non-small cell lung cancer (NSCLC) with those of 178,431 patients with de novo NSCLC [57]. When compared with patients with de novo NSCLC, cHL survivors diagnosed with NSCLC had significantly inferior stage-specific overall survival rates.

Risk factors — The risk of lung cancer is substantially increased in patients with cHL treated with RT or chemotherapy, especially among those patients who smoke [12,18,36,38,42,58]. In one study, the risk of lung cancer was threefold higher in cHL survivors with a first-degree relative with lung cancer, compared with those without a positive family history [38]. In that study, there was a more than additive interaction between a positive family history of lung cancer and cHL treatment.

In patients who have received radiation, most of the tumors are found within or adjoining the irradiated fields [19]. Patients who receive higher radiation doses appear to be at greater risk. One study, for example, found a strong dose-response relationship between radiation dose and the risk of lung cancer at that site [58]. There was a relative risk of 9.6 for a dose of >9 Gy compared with <1 Gy, with a possible downturn in risk for doses of 15 Gy or greater [58]. The decrease in lung cancer risk at the highest doses has also been noted after RT for breast cancer [59].

Chemotherapy appears to play a definite role in the development of second lung cancers following treatment of cHL. In one mature series of second malignancies occurring after treatment in 5519 British patients with cHL, lung cancer risk was significantly increased after chemotherapy (relative risk [RR] 3.3), RT (RR 2.9), and mixed modality treatment (RR 4.3) [49]. Most other studies have found an increased risk of lung cancer with chemotherapy treatment [12,34,42,56,60], but some have not [18,36,58].

There are conflicting data on whether certain regimens, such as those including lomustine and or multiple alkylating agents, are more likely to increase the risk [12,58]. As an example, in two reports, the risk of lung cancer was significantly greater in those receiving MOPP chemotherapy as compared with those not receiving MOPP; further, the risk increased with the number of cycles of MOPP [61,62]. In one of these studies, the risk of lung cancer was not increased following chlorambucil, vinblastine, procarbazine, and prednisone (ChlVPP), implicating mechlorethamine rather than procarbazine [61]. In the other, both mechlorethamine and procarbazine were implicated [62].

Smoking appears to considerably increase the risk of lung cancer after chest RT (RR of 13 for ever-smokers compared with never-smokers in one series) [60]. The risk seems to be significantly increased in relation to the amount smoked after the diagnosis of cHL (but not the amount smoked before diagnosis) [58]. There appears to be a more than a multiplicative relation between smoking-related and radiation-related risks (ie, the RR for smoking plus RT was greater than would be obtained by multiplying the separate RRs for these exposures). Similar findings were noted in a case-control study in which the risk of lung cancer after treatment with both alkylating agents and RT was additive, while the risk of smoking was multiplicative (table 2) [62].

Screening — There is evidence that low dose computed tomography (CT) scanning for patients at high risk for lung cancer allows for the discovery of a higher percentage of early stage disease, although the role of screening still remains controversial [63,64]. Screening recommendations are presented in more detail separately. In one single institution retrospective study, long-term survival in patients who developed lung cancer after cHL was only seen when the second cancer was identified as an incidental finding, providing additional evidence to support screening in high risk patients [65]. (See "Approach to the adult survivor of classic Hodgkin lymphoma", section on 'Late complications'.)

Breast cancer — The relative risk of breast cancer in females treated for cHL has been only slightly increased to approximately 1.5 in most studies [12,15,16,18,34,36,42,50], but higher values have been reported [13,66-68]. The absolute excess risks for all ages is not large; approximately 1 to 2 per 10,000 cHL patients per year in most series [12,18,42] with a higher value (13 per 10,000 per year) being described in a later United States study [13].

However, among females treated before the age of 30 and followed for long durations, breast cancer is one of the most serious long-term sequelae of RT for cHL. The raised risk of breast cancer occurs late, typically 10 or more years after first treatment [18,66,67].

Risk factors — The three major risk factors for breast cancer are the age at treatment for cHL, the use (and amount) of mantle RT, and the effect of alkylating and other agents on ovarian function (table 3) [69,70].

●Age at treatment – The risk of breast cancer is inversely related to age at treatment. Among females treated before the age of 16, the relative risk of breast cancer has ranged from 17 to 458 [13,14,39,51,67,71,72]. There is a similar variability in the actuarial risk of developing breast cancer among females under age 20, ranging from 12 percent to as high as 42 percent [40,44,51,73].

The risk declines steeply with age at treatment. In one study, for example, the actuarial risk for developing breast cancer 25 years after therapy was 34, 22, and 3.5 percent for females who received RT at age <20, 20 to 29, and ≥30, respectively [73]. In general, there is little if any increase in relative risk in females treated at age 30 or later [13,14,42,67,71,73,74]. This inverse relation to age suggests that the breast epithelium is sensitive to the carcinogenic effect of radiation during a "window" of risk before its terminal differentiation.

●Mantle RT – The main cause of the large risk of breast cancer after cHL is mantle RT [4,75]. Almost all cases have been in or at the margin of the radiation field. In three studies with a total of 59 cases, for example, breast cancer was in the field or at the margin in 54 [39,51,67]. An association between radiation field size and breast cancer risk has been demonstrated in several studies [4,68,76].

The importance of RT and the role of dose were evaluated in a large matched case-control study of 3817 young females with cHL diagnosed at ≤30 years of age [77]. A radiation dose of ≥4 Gy delivered to the breast (a dose that is routinely administered for cHL) was associated with a 3.2-fold increased risk of breast cancer (95% CI 1.4-8.2), as compared with the risk in patients who received both <4 Gy and no alkylating agents. The risk increased to eightfold with a dose of >40 Gy. A separately published subgroup analysis came to similar conclusions [78].

●Other factors – Other factors that might alter the risk of developing breast cancer in cHL include the effect of MOPP chemotherapy [67,73], splenectomy [79], and splenic irradiation [35,39]. However, the importance of these factors is not well established and, to the degree that it induces premature menopause, MOPP might actually be protective, but this regimen has been largely replaced by ABVD, which is not associated with premature menopause. (See "Factors that modify breast cancer risk in women".)

A large nested case-control study, and a separately published subgroup analysis, evaluated the effect of several factors other than RT in patients developing breast cancer after the diagnosis of cHL [77,78]. Treatment with both chemotherapy (MOPP or MOPP-like) and RT was associated with a significantly lower risk of developing breast cancer than treatment with RT alone (RR 0.45, 95% CI 0.22-0.91). Premature menopause following chemotherapy and/or pelvic (ovarian) radiation also resulted in a lessened risk. Taken together, these results suggest that hormonal stimulation may promote tumorigenesis in patients with cHL after RT has produced an initiating event. These results also suggest that treatment with preventative agents such as tamoxifen might provide a strategy to reduce breast cancer incidence in high-risk patients. An ongoing randomized trial comparing low dose tamoxifen versus placebo is evaluating this question (NCT01196936).

There are few data available on the effects of reproductive history, hormone supplementation, or genotype on the RT-related risks [80]. The risk for developing breast cancer was nearly doubled in females treated for cHL who had a first-degree relative with breast cancer, compared with those without a positive family history [38]. However, one retrospective study investigated the potential contribution of pregnancy to the late development of breast cancer in 382 females with cHL [81]. Females receiving RT during pregnancy or within one month of pregnancy had an increased incidence of breast cancer compared with nonpregnant females and females receiving RT later than one month after completion of pregnancy (hazard ratio 22, 95% CI 6-91).

Clinical and pathologic characteristics — Several retrospective studies have described the clinical and pathologic features of patients who developed breast cancer after treatment for cHL. Patients with second cancers appear to be younger with a higher incidence of bilateral disease than females with primary breast cancer. The median age at diagnosis is 30 to 40 years of age [52,71,82], while 20 to 25 percent have bilateral disease [82,83].

Breast cancer that develops after treatment for cHL may display more aggressive biologic characteristics when compared with sporadic cases of breast cancer. There have been several small hospital-based case-control and case-series evaluating the pathologic features of breast cancer in cHL survivors [82-87].

The largest study was an analysis of the Surveillance, Epidemiology, and End Results (SEER) database which included 2645 female patients who were diagnosed with cHL before age 35, treated with RT as part of their initial therapy between 1973 and 2000, and survived at least five years from initial therapy [88]. When compared with the general population, these patients had a sixfold increased risk of developing invasive breast cancer. The majority was ductal adenocarcinoma (79 percent) and developed in the upper outer quadrant of the breast. When compared with sporadic breast cancers, these tumors were more likely to display features associated with poor outcome such as estrogen and progesterone receptor negativity and high grade. Other studies have not seen these differences.

Prognosis — The prognosis of patients who develop breast cancer after treatment of cHL is variable, but appears to be worse overall than that of patients who develop breast cancer de novo [89]. However, because of the heightened awareness of the breast cancer risk, some studies have shown that these tumors are picked up at an early stage and have a good prognosis despite the presence of poor prognostic features. The prognosis is dependent on nodal status at presentation, as in patients with primary breast cancer. As an example, a study at Harvard Medical School evaluating MRI versus mammography has detected 16 breast cancers; only one patient had node-positive disease suggesting a relatively good prognosis for patients who develop breast cancer that is detected by careful surveillance after treatment of cHL.

In one series, for example, the five-year disease-free survival rates were much higher for patients with node-negative tumors compared with those with axillary involvement (85 versus 33 percent) [82]. In another report, the five-year disease-specific survival rate for pN0, pN1-3, and pN greater than 3, were 91, 66, and zero percent, respectively [83]. In addition, the tumor aggressiveness appears to vary. (See 'Clinical and pathologic characteristics' above.)

The majority of females who develop breast cancer and require local therapy to the breast are treated with mastectomy because of the risk of tissue necrosis after repeat irradiation of the breast [66]. With the use of partial breast irradiation with brachytherapy or conformal external beam irradiation, it may be possible to use lumpectomy and RT for patients with limited stage disease [90]. However, this approach should be offered as part of a clinical investigation.

Screening — A vigorous patient education program coupled with an early detection program for female patients treated with mantle irradiation might lead to earlier detection and improved survival [66,91-94]. One prospective trial for patients who were under 35 years of age at treatment demonstrated that the use of MRI, in addition to screening mammograms in this high risk population, identified additional breast cancers when compared with mammography alone [95].

Small prospective trials and retrospective analyses have demonstrated that screening programs can detect breast cancers in this population at an early stage. As an example, a prospective study of 90 female long-term survivors of cHL, all of whom had been treated with mantle irradiation, included patient questionnaires concerning awareness of the increased risk for breast cancer, written recommendations for breast examination and mammography, and annual follow-up [96]. Ten females developed 12 breast cancers during the study. All were evident on mammography; 8 of the 10 invasive cancers were node negative.

A retrospective study evaluated the use of screening breast MRI in 91 females with a history of chest irradiation [97]. The median age at the time of primary cancer diagnosis and presumed age at mediastinal radiation was 24 years. A total of 247 screening MRI examinations were performed over nine years. Suspicious lesions were detected on 27 MRI images (11 percent) among 21 patients, 30 of which were biopsied. Ten breast cancers were diagnosed; of these, four were detected by MRI alone, three were identifiable on MRI and mammography, while three were detected with mammography alone. These results suggest that MRI may provide additional information to that obtained by mammography but cannot be used in place of screening mammography in this population.

A prospective trial at Harvard Medical School performed annual breast MRI and mammogram over a three-year period in 148 females who had been treated with chest irradiation for cHL at age 35 years or younger and were greater than eight years beyond therapy [95]. Of 63 biopsies performed in 45 females, 18 (29 percent) detected a malignancy, all but one of which were preinvasive or subcentimeter node-negative disease. MRI detected five breast cancers that were missed by mammography, while mammography detected six breast cancers missed by MRI. MRI was associated with a higher false positive rate and led to more negative biopsies than mammography. Screening with MRI, mammogram, or both was associated with sensitivities of 67, 68, and 94 percent and specificities of 94, 93, and 90 percent, respectively, for the detection of breast cancer.

Outside of a clinical trial, our practice is to perform annual mammography in females treated with mantle irradiation starting 8 to 10 years after treatment or at age 40, whichever is first [98]. For patients who were age 30 or younger at treatment, we recommend an annual breast MRI as an adjunct to mammography. The American Cancer Society and American College of Radiology and the Society for Breast Imaging have recommended MRI screening in cHL survivors at high risk for breast cancer, such as those who received radiation to the chest between the ages of 10 and 35 years (table 3) [99-101]. The combination of mammography plus ultrasound may be an acceptable, less costly, alternative in this age group [102]. Either of these more intensive screening methods would be expected to increase the rate of false positives. However, earlier detection by such intense screening may also allow a subgroup to avoid cytotoxic drugs and further exposure to radiation. The role of chemoprevention with tamoxifen or raloxifene is yet undefined in this patient group. (See "Approach to the adult survivor of classic Hodgkin lymphoma", section on 'Late complications'.)

Bone and soft tissue cancer — Bone and soft tissue cancers, including nonepithelial malignancies of the breast, are rare. Thus, although patients treated for cHL have a large relative risk of developing these lesions [15,16,42,50], they account for only 5 to 10 percent of the absolute excess risk of second malignancy [12,13,15]. There are too few cases in most studies to assess variation in risk over time. In children treated for cancer generally, relative risks of bone cancer continue to rise from 2 to 4, to over 20 years after first treatment, with a relative risk of several hundred at 20 and more years follow-up [103]. (See "Breast sarcoma: Epidemiology, risk factors, clinical presentation, diagnosis, and staging".)

Bone and soft tissue tumor risks are greatly increased after RT [34,103]. As an example, a study in childhood cancer survivors found that the risk of bone sarcoma rose sharply with increasing radiation dose above 10 Gy, reaching a 40-fold excess at doses of 60 Gy or more [103].

There is some evidence that chemotherapy also may play a role in the development of bone and soft tissue tumors after cHL. One large study found a sixfold relative risk in relation to chemotherapy, with significant risks at zero to 4, 5 to 9, and 10 and more years after first treatment [34]. The increase in risk was associated with the administration of procarbazine, vincristine, doxorubicin, and bleomycin. In addition, bone sarcomas have been associated with alkylating agent administration in childhood cancer survivors in general [103].

Thyroid cancer — Thyroid cancers are uncommon after treatment of cHL. The absolute excess risk is approximately 1 to 3 per 10,000 patients per year, representing approximately 2 percent of the excess risk of second cancers overall [12,42,54]. The risk is much greater in females and after treatment in childhood [39,51,52,104]. A relative risk as high as 790 has been described in children treated between the ages of zero and four years [51]. The risk of thyroid cancer appears to be increased throughout follow-up to 20 and more years [39,50].

RT is the main cause of thyroid cancers after cHL, since the thyroid is a highly radiosensitive organ, particularly in children [52,105]. A possible link to chemotherapy with lomustine and other alkylating agents has also been suggested [34].

Small series have shown that thyroid cancers following RT for cHL usually develop within the radiation field [51,54]. This is compatible with the experience with thyroid cancer after childhood radiation exposure in other circumstances; 68 percent arose within the radiation field [52]. The risk increases with radiation dose up to 10 Gy and then appears to level off or even decrease [52,105]. (See "Radiation-induced thyroid disease".)

The prognosis of patients who develop thyroid cancer after cHL is poorly understood. In one series of 1981 patients treated for cHL and followed for a median of 14 years, 28 patients (1.4 percent) developed thyroid cancer [104]. At a median follow-up of 3.5 years since thyroid cancer diagnosis, 26 patients were alive without evidence of residual disease, one was alive with metastatic disease six years after diagnosis, and one died of metastatic disease 3.6 years after diagnosis.

Melanoma — The relative risk of malignant melanoma after treatment of cHL is approximately twofold and appears to be the same in patients treated in childhood [39,51] or as adults [13-16,34,36,42,50]. The absolute excess rate is 2 to 4 per 10,000 cHL patients per year [13-15,42].

In one study that evaluated the time course of this risk in six patients, most lesions occurred in the first five years of follow-up [53]. Two of the patients had more than one primary melanoma, and five had biopsy or clinical evidence of dysplastic nevus syndrome, suggesting that patients with this syndrome may be at increased risk.

The factors responsible for the apparent increase in melanoma risk in these patients are uncertain. There has been some evidence of increased risk in association with radiotherapy [34] but this is not a relation that would be expected from the radiation literature generally. It is possible that the melanoma risk is due to immunosuppression rather than a carcinogenic effect of treatment.

Gastrointestinal cancers — Gastrointestinal (GI) cancers account for approximately 10 percent of the excess risk of cancer after cHL; the relative risk is approximately 2.5 with an absolute excess risk of approximately 6 per 10,000 cHL patients per year [14,55]. The risk first increases at five years after treatment [14,16] and may continue to rise through 20 years [55,66].

Much greater risks of esophageal [51], colorectal [39,51], and gastric cancers [13,39] occur in patients treated under age 20. The relative risk of GI cancer falls with increasing age at first treatment in adults [55]. The risk of developing colorectal cancer was doubled in cHL survivors with a first-degree relative with colorectal cancer, compared with those without a positive family history [38].

The type of therapy for cHL appears to influence the risk of developing secondary GI cancer. In one study of 5519 British patients with cHL, the risk of developing secondary GI cancers was highest after combined modality therapy (RR = 3.3), lower after RT alone (RR = 1.7), and nonsignificantly increased following chemotherapy alone [49]. Others report similar findings [55,106].

In another study of 19,882 cHL survivors diagnosed from 1953 to 2003, 89 patients were diagnosed with gastric cancer with estimated cumulative incidence of 0.39 percent (95% CI 0.28-0.50 percent) at 15 years and 0.92 percent (95% CI 0.70-1.13 percent) at 30 years [107]. The risk of gastric cancer increased with increasing gastric radiation dose and with increasing exposure to alkylating agents, particularly procarbazine and dacarbazine.

The risk of pancreatic cancer after cHL has been associated with radiation dose and cumulative doses of alkylating chemotherapy [108], and the risk of colorectal cancer has been associated with radiation dose and procarbazine dose [109].

The prognosis of cHL survivors who develop GI cancers appears to be worse than that of patients with an initial primary GI cancer. A retrospective study of the Surveillance, Epidemiology, and End Results (SEER) database compared the survival of 209 cHL survivors with secondary GI cancer (105 colon, 35 stomach, 30 pancreas, 21 rectum, and 18 esophagus) to that of 484,165 patients with first primary GI cancers [110]. While outcomes following localized stage colon cancer were similar to those with primary disease, cHL survivors who developed advanced stage colon cancer or stomach cancer had significantly worse survival than those with primary disease.

Mesothelioma — Patients with cHL appear to be at an increased risk for developing malignant mesothelioma. This was best demonstrated in a cohort study of 2567 patients from the Netherlands who survived at least five years after treatment for cHL [111]. After a median follow-up of 18.1 years, malignant mesothelioma had been diagnosed in eight males and five females with a median time from treatment of cHL to diagnosis of mesothelioma of 27.7 years. When compared with the general population, there was an almost 26-fold increased risk for mesothelioma among patients treated for cHL (standardized incidence ratio of 25.7) resulting in an absolute excess of 3.5 cases of mesothelioma per 10,000 patients per year. In an exploratory analysis, factors that appeared to increase this risk included female sex, first treatment at age less than 31 years, and the incorporation of RT.

Other solid cancers — There is relatively little evidence concerning the risk of other solid cancers among patients treated for cHL [34]. On the basis of radiation epidemiology, it is likely that the risk of salivary gland tumors, nonmelanoma skin cancers, and bladder cancer may be increased after RT for cHL.

ACUTE LEUKEMIA — Acute leukemia was the first secondary malignancy to be noted in patients treated for cHL because of its relatively early appearance following treatment, as well as the rarity of acute leukemia in the general population. The highest risks and the greatest number of cases occur between 5 and 10 years after initiation of treatment, usually with alkylating agents, such as MOPP chemotherapy and MOPP/ABVD hybrid regimens, which are less commonly used in the current treatment of cHL [112]. This increased risk persists after 10 years, although at lower levels [12,14,16,41,50,112,113].

The reported incidence of acute leukemia in large series has ranged from 1.3 percent at an average follow-up of 6.7 years (0.8 percent in the subset in continuous complete remission) [16] to 3.3 percent at 15 years [15]. Although the relative risk of leukemia in these patients is considerably increased compared with the general population, the relatively low cumulative risk and number of excess cases of acute leukemia has limited the study of specific risk factors.

Risk factors — Most leukemias that occur after the treatment of cHL are acute myeloid leukemia (AML) related to alkylating agent exposure [12,15,16,26,34,42,114,115]. The relative risk increase associated with large-field irradiation alone is smaller [12,14,15,18,112]. The actuarial risk of leukemia at 15 years ranges from 0 to 0.6 percent among those receiving only radiation therapy (RT) to as high as 16.5 percent at 15 years in heavily treated groups [15,18,39,116,117].

The relative risk of leukemia also varies with the reference group and the type of leukemia considered. The relative risks in chemotherapy-treated patients tend to be over 20-fold increased in cohort analyses that use population-based comparisons of all leukemia risk, over 50-fold increased with population-based estimates of AML risk, and less in case-control analyses in which the reference group usually consisted of patients treated with RT alone [13,15,18,39,42,51,114,115].

Chemotherapy — The use of alkylating agents is known to be associated with an increased risk of developing AML. The leukemia risk appears to be most related to total dose of alkylating agents (eg, mechlorethamine, cyclophosphamide) or nitrosoureas [12,15,18,19,26,34,39,42,114,115,118-120]. Anthracyclines (eg, doxorubicin) and epipodophyllotoxins (eg, teniposide, etoposide), both of which are topoisomerase II inhibitors, may also increase the risk of leukemia. (See "Acute myeloid leukemia: Pathogenesis", section on 'Chemotherapy-induced AML'.)

●ABVD – The 15-year cumulative risk of AML after ABVD (doxorubicin, bleomycin, vinblastine, dacarbazine) has been estimated to be less than 1 percent [14,18,117,121]. This is reflected in a decrease in leukemia risk among those treated for cHL after 1980 when treatment with MOPP/ABVD largely replaced MOPP-like combinations (2.1 versus 6.4 percent in one series) [18].

There is, however, also concern about the role of anthracyclines (eg, doxorubicin) and epipodophyllotoxins (eg, teniposide, etoposide), both of which are topoisomerase II inhibitors, in the risk of leukemia. Limited evidence suggests that doxorubicin and epipodophyllotoxins in combination with higher doses of alkylating agents (including nitrosoureas) may have a synergistic effect in the risk of AML [115,120,122]. (See "Acute myeloid leukemia: Cytogenetic abnormalities", section on 'Therapy-related myeloid neoplasms'.)

●BEACOPP – The risk of developing AML within five years of treatment with escalated BEACOPP (bleomycin, etoposide, doxorubicin, cyclophosphamide, procarbazine, and prednisone) is higher than that with standard BEACOPP or COPP/ABVD (2.5 versus 0.6 and 0.4 percent, respectively) [123].

In an analysis of 11,952 patients treated for newly diagnosed cHL within German Hodgkin Study Group trials between 1993 and 2009 and followed for a median of 72 months, 106 patients (0.9 percent) developed AML/MDS at a median time from cHL treatment of 31 months [124]. AML/MDS was more common among those who received four or more cycles of BEACOPP when compared with those who received fewer cycles or no BEACOPP (1.7 versus 0.7 versus 0.3 percent, respectively).

●Stanford V – AML does not appear to be increased among patients receiving the Stanford V regimen (doxorubicin, vinblastine, mechlorethamine, vincristine, bleomycin, etoposide, and prednisone plus radiation). No cases of AML/MDS or non-Hodgkin lymphomas have occurred among a total of 256 patients treated in one series [125].

Radiation therapy — Most studies have found a modest risk of acute leukemia, often not of statistical significance, associated with RT alone, in comparison to the general population [12,15,18,112,116,126]. There is suggestive evidence that the risk may be higher in patients receiving extensive RT (eg, total nodal irradiation or large fields including the pelvis), in which case a large amount of the bone marrow is in the treatment field and the radiation is being used more akin to a systemic agent [16,34,114]. In one report, for example, patients treated with RT alone, who received a total radiation dose to the bone marrow >20 Gy, had a risk of leukemia that was eightfold higher than among those who received <10 Gy [114]. (See "Acute myeloid leukemia: Pathogenesis", section on 'Ionizing radiation'.)

Chemotherapy plus RT — There are conflicting reports as to whether combined modality treatment does [16,127-131] or does not [12,15,18,39,42,114,115] produce greater risk of AML than chemotherapy alone. However, the preponderance of data do not support the notion that the combination of chemotherapy and RT, in particular, mantle irradiation, with or without upper abdominal irradiation, confers a higher risk of leukemia than chemotherapy alone. However, there is evidence that the use of total nodal irradiation or large field and pelvic irradiation does increase the risk of leukemia seen in combination with alkylating agent chemotherapy. There is no evidence to suggest that the leukemia risk increases with increased radiation dose [114].

Age — Age over 35 to 40 years at treatment has been reported to be a significant risk factor for leukemia in some studies [13,14,16,18,112], but not in others [15,26,42]. The higher cumulative risk of AML in patients diagnosed over age 35 to 40 may simply reflect the higher baseline (general population) incidence of the disease in older persons.

Thrombocytopenia — Prolonged thrombocytopenia has been associated with an increased risk of developing leukemia after chemotherapy for cHL [115]. Rather than reflecting a direct relationship, the thrombocytopenia in these patients may simply have been a marker of increased bioavailability of chemotherapy [115]. Alternatively, there is considerable evidence that prolonged thrombocytopenia is a marker for permanent stem cell damage and that such patients have an increased risk for developing a leukemic clone.

Treatment and prognosis — The treatment and prognosis of therapy-associated AML is presented separately. (See "Therapy-related myeloid neoplasms: Epidemiology, causes, evaluation, and diagnosis".)

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Management of Hodgkin lymphoma".)

SUMMARY

●Description – Most patients with classic Hodgkin lymphoma (cHL) achieve long-term survival without a recurrence of cHL. With this success, late complications of treatment have emerged as competing causes of death and morbidity among cHL survivors. Long-term survivors are at risk for developing cardiovascular disease, pulmonary disease, second cancers, and other late complications.

●Late mortality in cHL survivors – Most deaths in the first 5 to 10 years after treatment of cHL are associated with disease relapse; later deaths are more likely to be due to other causes, including second cancers. (See 'Effect of elapsed time' above.)

●General observations – The most common secondary cancers in cHL survivors are solid tumors, leukemias, and lymphomas. The time course of development varies with the type of cancer; as examples, leukemia after alkylator-based treatment peaks at five to nine years post-treatment, while solid tumors typically develop after ≥10 years. (See 'General observations' above.)

●Relative and absolute risks – Patients treated with chemotherapy and/or radiation therapy (RT) for cHL are at an increased risk for second cancers compared with the general population. The increased relative risk in cHL survivors is >2-fold for solid tumors, 3- to 35-fold for lymphomas, and 10- to 80-fold for leukemia. The high relative risks for lymphomas and leukemia correspond to modest absolute risks because the incidence of these cancers in the general population is low. (See 'Relative and absolute risks' above.)

●Risk factors – Factors that contribute to the increase of second cancers in cHL survivors include (see 'General observations' above):

•Chemotherapy agents and doses

•Radiation dose and field

•Patient age at treatment

•Time since cHL treatment

•Family history

•Smoking

●Prognosis – Prognosis varies with the type of second cancer and host characteristics, but outcomes are generally poor. (See 'Prognosis' above.)

●Secondary cancers

•Non-Hodgkin lymphoma (NHL) – Combined modality treatment is associated with a greater risk for NHL than either chemotherapy or RT alone. (See 'Non-Hodgkin lymphoma' above.)

•Solid tumors – The most common second cancers in cHL survivors are:

-Breast – Contributing factors include age at treatment for cHL, use of mantle RT, and effects of alkylating agents and other chemotherapy on ovarian function (table 3). (See 'Breast cancer' above.)

-Lung – Chemotherapy, RT, and smoking are important risk factors for secondary lung cancer. (See 'Lung cancer' above.)

-Gastrointestinal tract – Risks for esophageal, colorectal, and gastric cancers are increased, especially in patients who were treated at younger ages. (See 'Gastrointestinal cancers' above.)

-Others – Other second cancers in cHL survivors include malignancies of bone/soft tissue, thyroid, melanoma, and others.

•Acute leukemia – Most cases of secondary leukemia occur between 5 and 10 years after treatment. Risk factors include chemotherapy, RT, and younger age at treatment of cHL. The risk for secondary leukemia has declined with less use of alkylating agents. (See 'Acute leukemia' above.)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges the late Peter M Mauch, MD, who contributed to an earlier version of this topic review.